---

At a glance

Current package: v2.9.4. /paseo-multi was exercised with Paseo 0.11.2 on the listed Mac/Windows combinations. The existing package compatibility baseline is 0.10.2; this is not a claim that all ten skills were verified on 0.11.2. Multi validation · Package baseline.

One-line summary

Paseobility is an agent-installable Paseo slash-skill pack: you hand this GitHub repo URL to Codex, Claude, or a Paseo agent and have it install.

Once installed, you can pull frequently used Paseo patterns out like slash commands: browser use (web browser manipulation), multi-agent orchestration, agent tournaments, session briefs, project bootstrap, pre-install repo security checks, archiving inactive subagents, and cross-device artifact sharing.

With base Paseo alone you can combine built-in tools to do similar work. But letting the agent re-decide that combination every time is slow and produces inconsistent results, so the pack bundles frequently used patterns for immediate reuse.

This repo is a skill package for reproducibly composing Paseo's built-in tools. It ships a macOS/Linux bootstrap helper and a Windows PowerShell install helper.

---

/paseo-god — explicit task mode (v0)

Define observable completion criteria, use only the needed debugging/review/spike
procedure, and retain a separate TODO, evidence and resume point for each run.
Learning is proposed for review; native memory and skills are not automatically
rewritten. The optional Node helper creates/resumes records, not autonomous jobs.

./scripts/paseobility-init.sh --skill paseo-god --no-context

.\scripts\paseobility-install.ps1 -Skill paseo-god

/paseo-god Fix this bug, verify the original trigger, and save a resume point if blocked.

See the God guide for scope, local records and validation
limits. Existing version and platform reports retain their original scope.

2026-10-19 — proportional God mode costs (v2.9.4)

/paseo-god now sizes fixed per-task costs to the change. In one measured
project, Git and checks took seconds while the commit cycle was slow because
required documents were reread every task, a second agent rediscovered the same
context and slow checks repeated after each fix. A run now classifies the change
once and records the checks that class needs, reads mandatory documents once per
run, runs slow checks once per batch and hands context over through RUN.md.
Project rules are still followed; a measured disproportionate rule cost becomes
a learning candidate. Details: the God guide.

2026-10-15 fix release — Windows and remote Macs (v2.9.3)

Running the installed pack on a Windows PC and three remote Macs reproduced
three defects, now fixed with tests:

• /paseo-share list, latest and fetch no longer fail when one artifact
  stops validating (for example an .mp4 an older release accepted). The
  invalid entry is skipped with a warning; only validated metadata is used.
• The doctor reports a stopped Cua daemon as stopped. cua-driver 0.34.x
  answers status with exit 1 on stderr, which was read as an unsupported
  subcommand.
• /paseo-multi exec --workspace ID supports daemons that require a
  workspace per terminal (Paseo 0.11.2: "workspaceId is required"), and the
  daemon's reason is shown instead of a bare failure. Credential-bearing runs
  should be closed right after completion.

Verified by the macOS and Windows CI and live on VAX_COM (Windows),
MAC_COM, mac_home and MAC_MINI.

2026-10-13 follow-up — all nine existing skills (v2.9.2)

The first refresh changed cleanup, browser, Cua and design. This follow-up also
reviews and improves multi, orchestration, project, share and spyware-check:
complete remote-agent pagination, evidence-based worktree integration,
monorepo-aware context, classified push retries and current scanner commands.
The existing nine-skill scope excludes the separately added God workflow.
See the per-skill coverage and validation
for official sources, local checks and remaining runtime limits.

2026-10-13 update — dependency refresh (v2.9.1)

• Browser recipes pin [email protected] with [email protected].
  Node must satisfy ^20.19.0 || >=22.12.0. The wrapper now supports ESM and
  CommonJS; the recipes keep dynamic import. The keyless free browser remains
  Chromium 145 on macOS, separately from the licensed Chromium 154 channel.
• Fresh Cua installations pin 0.34.1 at immutable commit
  0c69d9a2c8abbb5051a2df4a23f0fa166d5385dc. Skill instructions follow its
  token-only action contract and distinguish display previews from action captures.
  The package installer still reuses existing working drivers; host upgrades
  require an explicit update request.
• Vercel's offline interface-guideline snapshot is refreshed to the 2026-10-05
  commit, including desktop overscroll guidance.
• On the local Apple Silicon Mac, the updated browser passed fixture input,
  click, result readback, three viewport captures and failure-path cleanup.
  The user-requested Cua 0.28.2 → 0.34.1 upgrade preserved OS grants; version,
  doctor and tool schemas were checked. This is not a new native GUI end-to-end
  pass or Windows/Linux runtime validation.

The same local v2.9.1 preparation also hardens /paseo-agent-cleanup: complete
paginated state, exact workspace ownership, permission/unknown-state guards,
pre/post checks and exact-ID client dispatch replace CLI display-list assumptions.
A workspace-only request no longer selects unrelated disposable agents. CI now
checks the catalog/frontmatter, vendored runtime/data and all skill PowerShell
files. Upstream-only tests are explicitly skipped with reasons. Browser recipes
use accessible locators and wait for postconditions. See
the improvement validation.

See the dated validation record.
Earlier update sections retain their historical versions and evidence.

2026-10-11 update — /paseo-multi (v2.9.0)

/paseo-multi selects named computers connected through Paseo for status and
agent checks, remote Bash/PowerShell commands, and coordinated updates. It
checks the live server identity before operating, preserves active work before
restart, and captures remote command exit codes before reporting completion.
Trusted host connections live in a private local registry outside this package;
new Desktop connections require registration rather than automatic discovery.

The adapter was exercised from macOS and Windows controllers with the Paseo
0.11.2 client: each ran read-only commands on Mac and Windows targets, recovered
recorded runs and confirmed their intended nonzero exit codes. Twelve offline
adapter tests passed on both controllers; the shared Python suite passed all
49 tests on macOS. Run records are saved before dispatch, Windows JSON uses
UTF-8, and oversized Windows commands are rejected before submission. Linux
remains unverified, and captured Korean text may contain extra screen spaces.
See the dated validation report. Existing
compatibility records below retain their dated scope.

Install only this skill:

./scripts/paseobility-init.sh --skill paseo-multi --no-context

.\scripts\paseobility-install.ps1 -Skill paseo-multi

Then register trusted connections using the
connection guide. Installation
itself does not import a host registry or operate/restart a PC. First-use prompt:

/paseo-multi Check the registered PCs and their running agents.

2026-10-05 update — /paseo-design with design-check (v2.8.3)

v2.8.2 added /paseo-design, the web and app frontend workflow. v2.8.3 reworks
it around proof instead of prose:

• Project rules first. A repo with its own design skill, design document,
  token file, or UI checks is followed as-is; the skill never writes a second
  DESIGN.md or palette next to them.
• scripts/design-check.mjs renders a URL or HTML file at 375, 768 and
  1440 px in headless Chrome (Chrome, Edge, Chromium, or a Playwright download;
  no npm dependencies) and measures the page: overflow, clipped or overlapping
  text, left-edge and row misalignment down to 1 px, uneven gaps, card heights,
  WCAG contrast, tap targets, Hangul words split at line ends, Korean typography
  (font order, tracking, line length), spacing and type-scale sprawl, and the
  common AI tells (gradients, glass, accent rails, eyebrows, icon toppers,
  centered stacks), plus image aspect ratios, keyboard focus rings, animations
  that keep running under reduced motion, text sitting off-center in buttons,
  mixed icon sizes and control heights; open shadow roots are measured too, and
  --scheme light,dark renders both themes. It writes clean and marked
  screenshots, a contact sheet, a
  Markdown report with a fix hint per finding, and JSON; exit 1 while errors
  remain. --baseline ratchets legacy pages; --selftest proves the checker on
  the host.
• Compose, then check. The agent names one of seven surfaces before choosing
  a layout, follows references/design-digest.md (Anthropic frontend-design and
  Hermes claude-design condensed), reviews the screenshots against
  references/review-checklist.md, and reports errors 0 before calling UI work
  done.
• templates/DESIGN.md passes designmd lint (the dotted design.md bin is a
  silent no-op on Windows; the skill pins the working form). Vendored guides keep
  their upstream text and licenses (vendor/SOURCES.md, now with pinned
  commits); Vercel's Web Interface Guidelines ship as an offline snapshot.
• CI runs the design-check selftest on macOS and Windows.
• v2.8.4: content inside a visually hidden (screen-reader-only) ancestor, such as a
  table header kept for screen readers on phones, no longer counts as overlapping
  text. Found in an A/B run of the skill; the clean fixture now covers it.
  Unit charts, number pads, right-anchored labels, buttons in a bar and tilted
  illustrations no longer raise false alignment or nested-card findings.
• v2.8.4 also loosens the taste rules after real use showed over-restraint makes
  pages bland: the skill now starts from one concept taken from the subject's
  world, sketches three directions for new screens, treats the AI tells as habits
  (ai-* findings are info, not warnings) and judges blandness like breakage.

The 2026-09-30 CloakBrowser notes below stay as the previous update. Older
sections are historical and unchanged.

2026-09-30 update — CloakBrowser browser default, Paseo 0.10.2, Cua pin 0.30.4

This section records cumulative changes to the then 7-skill package.

/paseo-browser now defaults to CloakBrowser

• Default backend is CloakBrowser — a stealth Chromium wrapped by the pinned
  cloakbrowser package's Playwright API ([email protected],
  [email protected], Node >= 20). Ordinary navigation, reading, click/fill,
  bounded waits, read-only JS, screenshots, and console/network diagnostics are
  directly executable; recipes close the browser in finally so failures do not
  leak it.
• The pinned wrapper loads via dynamic import() from a Paseobility-owned
  runtime. Paseo browser_* and the Microsoft Playwright CLI remain optional,
  explicit backends (each reference file carries its backend label) and are
  never silently substituted when CloakBrowser fails.
• Skill copying stays docs-only: the installers (including a temp
  --target-home) do not install Node packages or download a browser. The
  runtime is set up explicitly:
  npm install --save-exact [email protected] [email protected] under a
  runtime dir, with CLOAKBROWSER_CACHE_DIR scoped to it.
• Tested locally (headless, keyless free binary): the fixture adder page was read,
  #a=7, #b=5, #compute produced 12, a fresh semantic snapshot confirmed
  it, and desktop (1280×720) plus mobile (375×812) screenshots were captured.
  Error paths (failed navigation; post-launch screenshot failure) exit non-zero
  and still close the browser. The binary launched on macOS was Chromium
  145.0.7632.109 (keyless free build); the latest keyed build (macOS 151,
  Linux/Windows 152) is free for one concurrent session with a GitHub-issued
  key or paid for scaling, and was not tested here.
• Limits: no verified external anti-bot/stealth result, no headed run, and no
  Windows/Linux runtime check in this update. macOS only.

Paseo 0.10.2 contract check (app/source), live daemon unchanged

• Installed app/CLI 0.10.2 matches the latest official v0.10.2 release; the
  running daemon was not restarted and still reports 0.9.0-beta.2.
• 0.10.2 is verified at the release-notes plus installed-app app.asar
TOOL_SPECS layer (61 tools, AGENT_FIELDS without a profile argument), not
  by a fresh 0.10.2 daemon handshake. See
  docs/compatibility-0.10.2.md.

Cua Driver source pin 0.30.4 (host stays 0.28.2)

• The Paseobility updater now pins Cua Driver 0.30.4 (immutable commit
  bf6c76786d938070f4ecf1e44004752f69f518b8), while the host keeps its working
  0.28.2 — an existing working driver is reused and never auto-upgraded.
• Source-reviewed installer behavior changes only: baked version, a
  withdrawn-release guard (0.28.3), and an autostart report fix;
  --no-modify-path / -NoPathUpdate, no-daemon-start, and Windows staging/
  rollback behavior are preserved. No new 0.30.4 GUI validation is claimed.
• The prior Cua preview records and tests are unchanged (offline suites still
  pass; no invented success).

Validation for this update is local and offline: skill frontmatter, diff checks,
offline scanner/shell suites, and a temp-home single-skill install. Windows runs
in GitHub Actions.

---

v2.8.1 — Windows Cua install fix, read-only doctor, offline tests

v2.8.1 fixes a Windows-only Cua Driver install conflict and adds a read-only
diagnostics helper plus offline tests. It ships no Cua or Paseo binary change.

• Windows Cua staging fix (code-level). scripts/paseobility-cua-driver.ps1
  stages into a private bin, validates the staged binary, publishes only the
  executable set, and rolls back only the files it created on failure. A fresh
  real-host Windows install was not reproduced; the fix is covered offline
  only (platform validation).
• Read-only doctor. scripts/paseobility-doctor.py (Python 3 stdlib) with
  thin .sh/.ps1 wrappers reports skill source/install integrity, Paseo
  version/reachability, and Cua version/permissions/daemon as distinct statuses;
  --json carries only allow-listed fields and --target-home isolates a check.
  --check-mcp is explicit-only and reports protocol/tools discovery, not a GUI
  result.
• Offline tests and CI. tests/test_doctor.py, tests/test_e2e_assets.py,
  and native tests/test_cua_driver_runtime.ps1 run on macOS and Windows CI; the
  installer passes its selected --target-home through to the doctor. Both
  doctor wrappers need Python 3, but diagnostics stay optional for installation.
• E2E assets and recovery reference. e2e/ holds a static browser fixture
  plus a foreground fixture server for a manual runbook (no automated report
  validator). skills/paseo-cua/references/recovery.md documents bounded retry,
  the permission flow, and that verify_state: unknown is not a pass.

See Doctor and tests and
Current release testing.

---

v2.8.0 — Added the Cua Driver skill

v2.8.0 adds paseo-cua. It is used only when the user explicitly asks to drive a native desktop app with the trycua Cua Driver; ordinary web UIs belong to paseo-browser, and ordinary coding/shell work belongs to the other skills. The package installer auto-ensures the driver for paseo-cua or the full package. The skill itself performs no installation: if it is copied manually without the runtime, it reports the prerequisite and a recovery instruction rather than promising an install.

Cua Driver runtime auto-install

Selecting paseo-cua, or installing the full package, also prepares the trycua Cua Driver runtime. If a working driver already exists, the installer reuses it and never auto-upgrades it.

• Pinned source: trycua/cua commit 0c69d9a2c8abbb5051a2df4a23f0fa166d5385dc (the cua-driver-rs-v0.34.1 release tag), Cua Driver version 0.34.1. This pin is a source/schema-reviewed contract; the last runtime exercised end-to-end in the validation records below remains 0.28.2, and an existing working driver is reused, never auto-upgraded.
• The installer downloads the upstream install scripts at the pinned commit and runs them from a temporary directory (never curl | bash). The pinned _install-rust.sh and _install-common.sh (and the Windows install.ps1 / _install-common.psm1) are fetched too, so the delegated script cannot silently fall back to a rolling URL.
• --skip-cua-driver / -SkipCuaDriver: skip the runtime and copy skills only (docs-only / offline).
• A custom --target-home / -TargetHome skips the real-host runtime by default and logs why. --allow-host-runtime / -AllowHostRuntime instead allows real host runtime installation even when the skills --target-home is custom. The driver is always installed at its normal host location; on macOS that still writes /Applications/CuaDriver.app and ~/.cua-driver, so this is not a sandbox.
• The driver step verifies the code signature on macOS, does not modify PATH (--no-modify-path / -NoPathUpdate), and never touches shell rc or MCP config. The pinned installer only resolves/downloads a release and stops stale daemons; it never starts a daemon (install.sh, _install-rust.sh). An explicit CUA_DRIVER_RS_VERSION=0.34.1 pin outranks the installer's baked version, and the pinned installer refuses withdrawn releases (0.28.3).
• If the runtime step fails, the installer exits non-zero and reports that skills were copied but the runtime setup failed (installation may be incomplete).
• macOS Accessibility and Screen Recording permissions are granted by the human. A binary being installed is not the same as permissions being ready.
• The Cua Driver ships product telemetry enabled by default (the installer does not change it). You can inspect or disable it yourself with cua-driver telemetry status / cua-driver telemetry disable; Paseobility never changes that setting. Collection details were not independently audited — see the platform validation status.
• A manual cp -R / Copy-Item copies documents only and cannot auto-install the Cua Driver runtime.

The package now has 10 skills. /paseo-god adds an explicit task workflow; its v0 validation is recorded separately in the God guide. CLI/MCP tool-schema compatibility was checked for the prior six skills at v2.7.0 and is recorded in the linked compatibility report, which does not cover paseo-cua, paseo-design, paseo-multi or paseo-god — the six-skill record must not be read as covering the later skills. paseo-cua is preview / limited validation, not broad stable Mac and Windows support. Its only directly verified boundary is a single Apple Silicon macOS 26.6.2 host on 2026-09-21 with Cua Driver 0.28.2; separate user-reported passes on Intel macOS and Windows, plus an isolated upstream PR build, are recorded with a dated matrix in the Cua platform validation status. See Verification status below for the compact 2026-09-22 summary.

Skill	Scope
/paseo-god	Explicit task mode: criteria, selective procedures, verification and run resumption
/paseo-multi	Selected remote computers: status, agents, shell commands and updates
/paseo-orchestration	Explicitly requested multi-agent coordination or compare/tournament
/paseo-project	Requested project summary/handoff or initial setup/environment changes
/paseo-browser	Web UI manipulation/verification via CloakBrowser (Paseo browser + Playwright CLI optional)
/paseo-cua	Explicitly requested native app GUI driven by the trycua Cua Driver
/paseo-design	Web or app frontend: project rules first, compose surface-first, then prove it with design-check (screenshots plus measured findings, errors 0)
/paseo-agent-cleanup	Cleaning up selected test agents/workspaces
/paseo-share	Artifact sharing between personal devices
/paseo-spyware-check	Pre-install static security inspection of a repo

Orchestration keeps allow_implicit_invocation: false. The project skill does not run for ordinary coding or session resume alone. Only the reference for the selected mode is read for compare/coordinator and summary/setup details.

Former skill	Merged into
paseo-agent-tournament	paseo-orchestration comparison mode
paseo-session-brief	paseo-project read-only brief mode
paseo-project-bootstrap	paseo-project setup mode
paseo-computer-use	paseo-browser
paseo-skill-save	Removed; personal library data preserved

When updating an existing install, the following command moves pre-consolidation skills to backup. Use --with-claude / -WithClaude only when you also want to update the Claude install.

./scripts/paseobility-init.sh --migrate-skills --no-context

.\scripts\paseobility-install.ps1 -MigrateSkills

An ordinary copy/install does not remove old names. A selected install moves only that skill's predecessor names. Backups of predecessor names are kept even with --no-backup / -NoBackup. Unrelated skills and private runtime are untouched, and installation aborts if a predecessor directory's name/ownership cannot be confirmed or if it is a link.

---

v2.6.0 update — Paseo 0.6.1 compatibility

v2.6.0 re-checked 8 features against the then-current Paseo 0.6.1 CLI and built-in tool schemas, and updated the features that connect directly to the Paseo runtime.

• orchestration/tournament dropped the legacy paseo_* tool names and now prefer list_profiles, use profile materialization, and follow settings.modeId, settings.thinkingOptionId, and settings.features.
• It does not guess provider/model; it confirms with list_providers, inspect_provider, and list_models only when needed.
• Worktree workspaces, supervised workspace scripts, agent notifications, the full schedule lifecycle, and heartbeat delete/recreate semantics were reflected.
• paseo-browser reflects the canonical browser_* tool names, workspace prerequisites, the connected desktop browser host, and the latest stale-ref/error recovery flow.
• The old /paseo-computer-use could be mistaken for full OS control, so it was renamed /paseo-browser. The installer does not auto-remove the old installed name, so an existing paseo-computer-use directory must be removed separately after an update.
• agent-cleanup validates the latest archive JSON { agentId|workspaceId, status: "archived", archivedAt }. It no longer requires the removed providerRelease field, so it does not misread a valid archive as a partial failure, and it protects initializing agents as active.
• project-bootstrap/session-brief read Paseo version/daemon/project/workspace and paseo.json workspace script state read-only.
• share/spyware-check state their compatibility boundary as local helpers independent of the daemon/MCP. The spyware receipt preserves external API and credential names declared in document-style skills as capabilities distinct from executable code, keeping the review gate.

Compatibility references: Paseo CLI, MCP tools, workspaces/worktrees, schedules, and local Paseo 0.6.1.

---

v2.5.2 update

v2.5.2 fixes a safety problem where /paseo-agent-cleanup archived healthy idle Codex sessions based on status alone.

• A run with no arguments is now a dry-run and does not use the default select-everything .*.
• Unfiltered --auto now targets only inactive agents with clear disposable/test/validation markers. Ordinary idle sessions are preserved as healthy sessions that can still be used.
• Scope can be limited with an explicit --agent <id> or a user-specified --pattern, and active agents and unapproved workspaces remain protected.
• After archiving, it verifies both the latest JSON acknowledgement and that the Paseo record disappeared from the active list.
• During verification it never opens history/timeline/resume, and never performs delete, stop, kill, lock-file deletion, or daemon restart.

v2.5.1 update

v2.5.1 fixes an agent UX regression in /paseo-share.

• /paseo-share: top-level help and --help print usage and exit 0 with no configuration, Git, GitHub, or filesystem side effects. Unknown commands remain non-zero as before.

v2.4.1 update

v2.4.1 fixes a bug where the /paseo-spyware-check macOS/Linux helper exited successfully when it could not create the report directory or required output files. Required I/O failures now exit non-zero immediately, while optional external scanner failures are recorded in tools.log and the remaining checks plus the fallback static scan continue.

v2.4.0 update

v2.4.0 had /paseo-share safely auto-prepare the authenticated account's private paseo_share repository. At that time /paseo-agent-cleanup auto-archived every inactive agent, but this behavior was replaced in v2.5.2 by the safer dry-run/explicit-selection policy.

Feature added/reinforced	Role
/paseo-share	Publish files to the private paseo_share and return a share ID plus preview/download links. Other computers fetch after verification
/paseo-spyware-check	Before installing a GitHub URL or local repo, read-only inspection for malicious install scripts, secret access, remote code execution, exfiltration, and supply-chain risk signals
/paseo-agent-cleanup	Preserve ordinary idle sessions and archive only explicitly selected or disposable-marked inactive agents. Workspaces are cleaned archive-only after approval
Installer backup	Automatically back up an existing same-name skill directory under skills-backups/ before overwriting
Report summary	Add High / Medium / Info counts and a verdict hint to the spyware helper report

/paseo-spyware-check calls external open-source scanner CLIs installed locally. Paseobility does not bundle or redistribute those scanners' binaries or rule sets.

/paseo-agent-cleanup is a dry-run by default and only shows candidates. It preserves ordinary idle sessions and archives only inactive agents selected by an explicit ID, a user-specified pattern, or a clear disposable/test/validation marker. Active agents such as running ones are never touched, and workspace cleanup is archive-only after approval.

/paseo-share checks GitHub CLI authentication on the first real share request and prepares <login>/paseo_share as a private repository. Onboarding from another computer with the same GitHub account reuses the existing repository after a safety check. Forgejo or another repository can be connected with an explicit setup <repo-url>. Uploaded files can be previewed directly on mobile via a link, and Paseo on another computer fetches them automatically using the share ID.

---

How to use

The primary user of this project is not someone installing from a shell directly, but someone who gives an AI agent the GitHub URL and delegates installation and verification.

To install only paseo-share, send the following sentence verbatim to Codex or Claude.

https://github.com/wilgon456/Paseobility

Read the instructions in this repository, install only paseo-share on this computer, and verify it.
Back up any existing install and do not restart the Paseo daemon.
When it is done, tell me how to start a new session/reload and how to make the first share request.

https://github.com/wilgon456/Paseobility

Read this repo and install into my local Paseo skills directory following AGENTS.md.
First test the install with a temporary --target-home/-TargetHome path, and install into the real skills directory once it passes.
After installation, also check that /paseo-project is recognized in a new Paseo session.

The agent reads AGENTS.md in Codex/Paseo and CLAUDE.md in Claude Code, then follows the OS-specific paths and install procedure. This Paseobility repository is the install source. Actual shared artifacts live in the private paseo_share repository of the authenticated account; installing never creates a repository by itself.

When you explicitly request these tasks, you can use the following capabilities:

• Open, read, click, type into, and verify web pages with screenshots.
• Split work across multiple agents and synthesize the results.
• Compare answers from several models such as GPT/Claude/DeepSeek/Grok and pick a winner or a merged plan.
• Before installing a GitHub URL or local repo, read-only inspection of spyware/supply-chain risk signals.
• Archive inactive agents with test-only markers after confirming scope, preserve ordinary idle sessions, and archive workspaces after approval.
• Publish work artifacts to a private Git share and open or fetch them from another computer or mobile.
• On request, summarize repo context, commands, instructions, and risks on one page.
• When initial setup is requested, gather README, docs, Claude/Codex/Cursor-style instructions to build working context.
• Design cross-provider collaboration such as implementing with Codex and reviewing with Claude.
• Put clear guardrails on infinite loops, dangerous submissions, and account changes.

---

Included skills

Skill	Role	Strong at requests like
/paseo-god	Explicit task execution and resumption	Bounded work, acceptance evidence, debugging/review/spike and learning proposals
/paseo-multi	Operate selected Paseo computers	Named host selection, remote status/agents, Bash/PowerShell commands and verified updates
/paseo-design	Web and app frontend workflow	Project design rules, existing components and frontend review
/paseo-browser	Browser manipulation workflow (CloakBrowser default)	Filling login forms, reading search results, clicking UI, responsive screenshots, checking web app state; CloakBrowser stealth Chromium via Playwright API, with optional Paseo browser_* and Microsoft Playwright CLI backends
/paseo-cua	Native desktop GUI driving	App window manipulation/snapshot/verification with the trycua Cua Driver, only on explicit request
/paseo-orchestration	Multi-agent coordination and comparison	Explicitly requested coordination or tournament, mode-specific guidance
/paseo-project	Project summary and setup	Requested read-only brief or environment/context setup
/paseo-share	Computer/mobile artifact sharing	Auto-prepare private paseo_share, clickable preview/download links, share-ID-based verification and fetch
/paseo-spyware-check	Pre-install security/spyware static check	Checking GitHub URLs, local repos, install scripts, secrets, exfiltration, supply-chain risk
/paseo-agent-cleanup	Inactive agent/workspace cleanup	Dry-run by default, preserve ordinary idle, archive only explicit/test-marked candidates, protect active agents, archive workspaces after approval

---

Install

The recommended install method is to give an AI agent this repository URL and let it install.

https://github.com/wilgon456/Paseobility
Read this repo and install into my local Paseo skills directory.
Copy skills/* for my Windows/Mac environment and tell me how to reload afterward.

The agent checks the Mac/Windows skills paths based on AGENTS.md in Codex/Paseo and CLAUDE.md in Claude Code. A paseo-share request installs only that skill instead of the full pack.

You can also install manually.

git clone https://github.com/wilgon456/Paseobility.git
cd Paseobility

# Install the Paseo / Codex skills only
./scripts/paseobility-init.sh --no-context

# Update just one skill
./scripts/paseobility-init.sh --skill paseo-agent-cleanup --no-context
./scripts/paseobility-init.sh --skill paseo-spyware-check --no-context
./scripts/paseobility-init.sh --skill paseo-share --no-context

# Also install for Claude Code
./scripts/paseobility-init.sh --with-claude --no-context

# Bootstrap a specific project right away
./scripts/paseobility-init.sh --root /path/to/your/project

On Windows PowerShell:

git clone https://github.com/wilgon456/Paseobility.git
cd Paseobility

# Install the Paseo / Codex skills
.\scripts\paseobility-install.ps1

# Update just one skill
.\scripts\paseobility-install.ps1 -Skill paseo-agent-cleanup
.\scripts\paseobility-install.ps1 -Skill paseo-spyware-check
.\scripts\paseobility-install.ps1 -Skill paseo-share

# Test-install into a temporary home first
.\scripts\paseobility-install.ps1 -TargetHome $tmp.FullName

# Also install for Claude Code
.\scripts\paseobility-install.ps1 -WithClaude

Cua Driver runtime

Selecting paseo-cua, or installing the full package, also ensures the trycua Cua Driver runtime. Other skill selections have no driver side effects. The runtime step is idempotent: an existing working driver is reused and never auto-upgraded.

# paseo-cua selected -> driver auto-ensured
./scripts/paseobility-init.sh --skill paseo-cua --no-context

# Skills only, no runtime (docs-only / offline)
./scripts/paseobility-init.sh --skill paseo-cua --skip-cua-driver --no-context

# Custom target homes skip the real-host runtime by default and log why
./scripts/paseobility-init.sh --target-home "$tmp" --no-context

# Explicitly allow real host runtime install with a custom target-home
./scripts/paseobility-init.sh --skill paseo-cua --target-home "$tmp" \
  --allow-host-runtime --no-context

.\scripts\paseobility-install.ps1 -Skill paseo-cua
.\scripts\paseobility-install.ps1 -Skill paseo-cua -SkipCuaDriver
.\scripts\paseobility-install.ps1 -TargetHome $tmp.FullName
.\scripts\paseobility-install.ps1 -Skill paseo-cua -TargetHome $tmp.FullName -AllowHostRuntime

The driver helper can also be run directly:

./scripts/paseobility-cua-driver.sh --check       # presence only (exit 3 when missing)
./scripts/paseobility-cua-driver.sh --dry-run     # print the plan, execute nothing
./scripts/paseobility-cua-driver.sh               # ensure the pinned driver

If the runtime step fails, the installer exits non-zero and reports that skills were copied but the runtime setup failed (installation may be incomplete); re-run or pass --skip-cua-driver / -SkipCuaDriver to install skills only.

Doctor and tests

The doctor is read-only, bounded, and cross-platform (Python 3 stdlib). Text mode
is for a human; --json is the shareable artifact carrying only allow-listed
fields (statuses, counts, validated skill names, sanitized semantic versions) —
never raw command output. Exit code 0 means "a report was generated", not
"everything is ready". --source-root overrides the skill pack; --target-home
points at an isolated install root.

python3 scripts/paseobility-doctor.py --root /path/to/project
python3 scripts/paseobility-doctor.py --root /path/to/project --json
# Explicit-only MCP protocol/tools discovery (needs a running daemon; not a GUI test):
python3 scripts/paseobility-doctor.py --root /path/to/project --check-mcp
# Isolated install check against a temporary home:
python3 scripts/paseobility-doctor.py --target-home /tmp/isolated-home --json
# Thin wrapper with the same passthrough:
./scripts/paseobility-doctor.sh --root /path/to/project

.\scripts\paseobility-doctor.ps1 -Root .
# Same flags as the Python helper (-SourceRoot/-TargetHome/-CuaBin/-CuaBinDir):
.\scripts\paseobility-doctor.ps1 -SourceRoot . -TargetHome $tmp.FullName -Json

Offline test suites:

# macOS/Linux
python3 -m unittest discover -s tests -p "test_*.py" -v

# Windows (no network, injected fixture installer)
pwsh -NoProfile -File tests/test_cua_driver_runtime.ps1

The installer backs up an existing same-name skill before overwriting.

OS	Backup path
macOS/Linux	~/.agents/skills-backups/Paseobility-<version>-<timestamp>/
Windows	%USERPROFILE%\.agents\skills-backups\Paseobility-<version>-<timestamp>\

If you must force-replace without a backup, macOS/Linux uses --no-backup and Windows uses -NoBackup.

After installation, start a new agent in the Paseo app, or reload integrations/skills in Settings.

If you prefer a manual copy (documents only — it cannot auto-install the Cua Driver runtime):

mkdir -p ~/.agents/skills
cp -R skills/* ~/.agents/skills/

Windows manual install:

New-Item -ItemType Directory -Force "$env:USERPROFILE\.agents\skills" | Out-Null
Copy-Item -Recurse -Force ".\skills\*" "$env:USERPROFILE\.agents\skills\"

---

Requirements

• Paseo desktop app
• Paseo Settings -> Agents -> Enable Paseo tools enabled
• One or more AI providers
  • e.g. Codex, Claude Code, etc.
• /paseo-share needs Node.js and Git. Default GitHub auto-onboarding needs an authenticated GitHub CLI (gh); Forgejo and custom remotes connect with an explicit repository URL. GitHub Actions is not used.
• /paseo-cua needs the trycua Cua Driver. The installer auto-ensures it for paseo-cua or the full package; macOS Accessibility and Screen Recording grants stay manual.
• Orchestration prefers Paseo's list_profiles. The legacy ~/.paseo/orchestration-preferences.json is read only as optional extra guidance.
• Install target skills path
  • macOS/Linux: ~/.agents/skills
  • Windows: %USERPROFILE%\.agents\skills
• The bootstrap context script targets macOS/Linux shell environments
  • On Windows, scripts/paseobility-install.ps1 only assists with skill installation.

---

Verification status

The CLI/MCP compatibility record for the existing six skills is the v2.7.0
compatibility report, which does not cover paseo-cua.
The six-skill record must not be read as covering all seven skills. The current
package compatibility baseline is Paseo 0.10.2; its re-verification record
is docs/compatibility-0.10.2.md.

paseo-cua is preview / limited validation — not broad stable Mac and Windows support.
Current dated status (2026-09-22):

Environment	Native GUI (Cua)	Browser result and limits
Apple Silicon macOS 26.6.2 · Cua 0.28.2 · direct local	Calculator input/read-back + PNG pass; verify_state unknown	viewport pass; built-in fullPage fails
Intel Mac mini 16 GiB · macOS 15.8 · Paseo 0.8.0 · Cua 0.28.2	user-reported: native GUI + PNG pass; verify_state unknown	browser input/viewport pass; built-in fullPage fails; same-tab scroll+stitch whole-PNG passed for two DPR-1 fixtures only (not a general helper)
Intel MacBook 8 GiB · macOS 15.7.9 · Paseo 0.7.2 · Cua 0.17.0 (reused)	user-reported prior E2E: native GUI/PNG pass; verify_state unknown	browser input/viewport pass; built-in fullPage fails; installed 0.9.0-beta.2 left unchanged and not revalidated
MacBook follow-up · isolated upstream PR #3197 build (reported 0.3.1)	—	user-reported: real MCP browser_screenshot(fullPage:true) passed 2573px/3511px at DPR 2 twice each, repeat SHA identical
Windows 11 25H2 · Paseo 0.9.0-beta.2 · Cua 0.28.2	user-reported: MCP 57 tools + Calculator input/AX/PNG/verify_state pass	viewport/full-page PNG failed (screenshot_no_frame then 15s timeout, no image); DOM-fixture input/click/scroll pass; localhost HTTP failed; physical browser host unknown

Notes:

• Only the Apple Silicon row was directly verified in this review; the Intel Mac and Windows rows are user-reported and were not independently reproduced. The MacBook follow-up used the isolated upstream PR build without additional product edits and is not a fix released or installed by Paseobility — the 0.9 backport was aborted on conflicts, and a 0.3.1 downgrade is not recommended.
• Native passes cover narrow tested flows, not blanket platform support. The 8 GiB sample is not a RAM minimum or guarantee, and memory was not measured at peak.
• Windows automatic runtime installation had a code-level conflict: the shared non-junction .local/bin was pre-created before the pinned upstream installer, which refuses an existing non-junction directory. v2.8.1 installs into a private staging bin and publishes only the executable set, and tests/test_cua_driver_runtime.ps1 covers the conflict, rollback, and shared-bin preservation offline. A real-host fresh Windows install was not reproduced for this release, so the earlier manual hardlink recovery remains the last real-host evidence. Two explicit-only skills remain unconfirmed in the Windows catalog.
• Test counts are as reported and were not re-run for this README update: 41 automated tests on the Intel Mac mini (cleanup 11 + share 15 + scanner 13 + shell 2); Windows 63 passed / 2 failed / 1 skipped, unchanged.
• The PowerShell wrappers were source-reviewed only on macOS (pwsh unavailable); the supplied Windows report confirms skill migration and manual runtime recovery. Isolated-path install/migration and helper regression tests were confirmed.
• The v2.7 six-skill compatibility doc does not prove all seven skills at runtime. This documentation update does not modify the Paseo app or capture engine. Do not read binary presence as permission readiness. Other macOS versions and Linux remain unverified.

Current release testing (v2.8.1) — separate from the device rows above

The rows above are older device evidence. The v2.8.1 release ran its offline
suites on macOS and Windows CI; CI run
35708474356
(head 9932ea7590ef137552d8b2b4e065e7fbb1792254) is SUCCESS for both the
macOS and Windows jobs. Exact scope:

• macOS: 49 Python tests discovered, 49 pass, 0 skip (Cua Driver runtime
  18, doctor 21, E2E assets 3, skill migration 7), plus Node cleanup 11, Node
  share 15, Python scanner 13, and shell 2 — 90 automated tests
  total.
• Windows shared Python suite: 49 discovered, of which 14 pass and the
  other 35 are POSIX-only skips. The Windows job also runs Node cleanup
  11, Node share 15, and Python scanner 13, all passing.
• Native PowerShell fixture suite: runs under both pwsh and Windows
  PowerShell 5.1, each reporting 109 assertions, 0 fail, 0 skip. This counts
  assertions, not 109 test cases, and it is a fixture suite — not a real
  downloaded-Cua GUI test.
• Still unverified: a fresh native Windows Cua install from the official
  download path was not reproduced. The fixture suite proves the staged
  install / shared-bin preservation / rollback / skill-migration contract, not a
  live download.
• ./scripts/paseobility-doctor.sh --root . was run in text and --json modes
  against this checkout; the JSON is deterministic, carries only allow-listed
  fields, and reports the live host as Cua 0.28.2 (permissions granted, daemon
  running) and Paseo 0.9.0-beta.2 reachable.
• With an explicit --check-mcp, the live Cua 0.28.2 MCP probe was observed
  negotiating the protocol and listing 56 tools on one connection. The probe
  spawns its own stdio server child, and the 56 tools come from that
  parent-run probe. This is a protocol/tools check only, never a GUI result, and
  it does not prove a Paseo agent client's MCP config or tool discovery.
• GUI end-to-end is blocked, not passed. A manual attempt to drive the
  e2e/ browser fixture on this host failed: browser_new_tab http://127.0.0.1:<port>
  timed out waiting for tab registration, no connected tab appeared, and the
  fixture server was terminated with no listener left on the port. The current
  browser host could not register the tab, so no GUI pass is claimed. The
  historical device rows above are unchanged.
• The updated paseo-cua skill was installed locally for this release into both
  the Codex/Paseo and Claude skill directories, with backups; all seven source
  trees match, the other twelve installed trees were untouched, and the Cua
  runtime stayed at 0.28.2.

Rollback

• These changes are scripts, tests, docs, and E2E assets; no Cua or Paseo binary
  is modified. To restore the previous helper behavior, the v2.8.0 tree is
  referenced but is not a guaranteed tag; the safest option is to check out
  the previous main commit 970ea8cab7c3fef757d2bd13f58c8389758aad1d in a
  separate clone, preserving your existing work. Skill rollback uses the
  installer backups.
• There is no automatic runtime rollback. Paseobility never auto-upgrades the
  installed Cua Driver, so a driver that already worked is left as it was.
• Diagnostics need Python 3, but installation never does: both
  paseobility-doctor.sh and paseobility-doctor.ps1 require Python 3 and exit
  non-zero with a clear message when it is absent, and the installer treats that
  as "diagnostics unavailable" and still completes. Without Python 3, skip
  diagnostics.

Details, including the optional Playwright fallback and per-host probes, are in the
Cua platform validation status and the
compatibility report. The entries below are prior-version records and are not evidence for the current version.

Past verification record — separate from this consolidation

Environment	Status	What was checked
Paseo 0.6.1 on Windows	Tested locally	CLI/daemon 0.6.1 agreement, agent/workspace/provider JSON, worktree/schedule/heartbeat/archive CLI schema, cross-check against current MCP/profile/browser sources, PowerShell temporary install
Apple Silicon macOS	Tested	Paseo CLI 0.2.5 detected Darwin/arm64, temporary HOME install, real ~/.agents/skills install, context generation, package script detection, /paseo-session-brief recognized by a new Paseo agent
Intel macOS	Tested	Paseo CLI 0.2.5 detected Darwin/x86_64, temporary HOME install, real ~/.agents/skills install, context generation, package script detection, /paseo-session-brief recognized by a new Paseo agent
Windows	Tested	Windows 11 x64, Windows PowerShell 5.1, Paseo CLI 0.2.5: PowerShell installer, -TargetHome temporary install, real %USERPROFILE%\.agents\skills install, /paseo-session-brief recognized by a new Paseo agent. Native bash context script not verified
/paseo-spyware-check on Apple Silicon macOS	Tested	On Paseo 0.3.0: temp HOME install, helper script run, fixture risk-pattern detection, recognition by a new Paseo agent, scanner dry-run. Required report I/O fail-closed and optional scanner failure continuation regression tests passed
/paseo-spyware-check on Intel macOS	Tested	On Darwin x86_64 / Paseo 0.3.0: install, helper script run, fixture risk-pattern detection, recognition by a new Paseo agent, Gitleaks secret scan no finding
/paseo-spyware-check on Windows	Tested	On Windows 11 x64 / PowerShell 5.1 / Paseo 0.3.0: native PowerShell static-search workflow regex compile, fixture risk-pattern detection, recognition by a new Paseo agent. Bash helper not verified on native Windows
/paseo-agent-cleanup helper	Tested locally	Node regression tests covering bare dry-run, ordinary idle preservation, disposable marker/explicit ID/user-pattern restriction, running/initializing protection, Paseo 0.6 archive acknowledgement and post-archive list verification, workspace approval gate, no dangerous command execution
/paseo-agent-cleanup on Windows	Previous install verified	On Windows 10.0.26200 x64 / Node v24.14.0 / Paseo 0.3.0: %USERPROFILE%\.agents\skills install and dry-run JSON. v2.5.2 policy regression verified by cross-platform Node unit tests
/paseo-share on Apple Silicon macOS	Tested	Official skill validator, Node security/onboarding/help and --help regression tests, private repo create/reuse and public/irrelevant repo rejection, real private GitHub publish/read/auto-fetch/original SHA-256 comparison, isolated Codex/Claude install
/paseo-share on Windows	Tested	Windows private checkout, PowerShell install, private GitHub connection, real TXT publish and remote file read, mobile GitHub preview

Intel Mac testing observed no install failure, agent recognition failure, or Intel-specific error.
Windows testing passed the PowerShell installer, -TargetHome temporary home install, and real skill recognition. Temporary home installs isolate on -TargetHome or $env:USERPROFILE instead of $HOME.

---

Quick usage examples

Sharing artifacts with another computer and mobile

On the first share request, GitHub connectivity is checked and the authenticated
account's private paseo_share repository is prepared automatically.

/paseo-share Share this report so I can view it on my phone.

If GitHub authentication is missing, it first guides you to
gh auth login --hostname github.com. To use Forgejo or a separate repository,
use setup <repo-url> explicitly.

Then publish or fetch in natural language.

/paseo-share Share this report so I can view it on my phone.
/paseo-share Show me the file I just shared from another computer.
/paseo-share Fetch the latest shared file into the current project and summarize it.

The publish result includes a share ID and clickable preview/download links. Private repository links require GitHub/Forgejo login in the mobile browser.

Driving a web UI directly

/paseo-browser
Open the login page at https://example.com, inspect the form structure,
then verify with a screenshot whether a test account logs in.

A possible workflow:

Open a new tab -> page snapshot -> find input fields -> enter values
-> click the button -> wait -> verify again with snapshot/screenshot

Commanding several agents

/paseo-orchestration
Split this feature into implementation, testing, and review, run them in parallel,
and have a review agent decide pass/fail at the end.

A possible workflow:

Decompose task -> create worker agents -> run in parallel
-> collect results -> review gate -> pass/block decision

Comparing answers from several models

/paseo-orchestration
On this README direction, have GPT argue for it, OpenCode DeepSeek against it,
and Grok compare the two and produce a final judgment.

A possible workflow:

Define participant roles -> run agents in parallel
-> collect results -> create a judge agent
-> produce winner / merged plan / risks

Making a session brief

/paseo-project
Assume I'm seeing this project for the first time and produce a brief I can work from.

Output scope:

Project -> Current State -> Instructions -> Commands
-> Project Map -> Risks -> Suggested First Moves

Setting up new project context

/paseo-project
Assume I'm seeing this repo for the first time; read docs, CLAUDE.md, AGENTS.md,
and package scripts to build working context.

You can also prepare the same thing with shell scripts first.

./scripts/paseobility-doctor.sh
./scripts/paseobility-context.sh

Generated files:

.paseobility/
├── context.md        # summary of README/docs/instruction files
├── commands.md       # candidate install/dev/build/test commands
├── project-map.md    # map of key files/directories
└── bootstrap-log.md  # OS, arch, Paseo state, warnings

Pre-install spyware check

/paseo-spyware-check
Is it safe to install https://github.com/owner/repo?
Focus on install scripts, secret access, remote code execution, and data exfiltration risk.

Core principles:

• It reads the repo read-only and does not execute its code.
• On macOS/Linux it uses the bundled helper, if present, to make a static report after a temp clone.
• On Windows it can make a native static report with the bundled PowerShell helper.
• If gitleaks, trufflehog, semgrep, osv-scanner, trivy, shellcheck, or yara are installed it uses them; otherwise it falls back to rg-based heuristics.
• If no scanner is present, the bundled helper first shows the install command, and can install via Homebrew after approval.
• Results are classified by severity as High, Medium, and Info, and self-references inside scanner docs/regex are marked Info.
• The helper report includes a Finding Summary that first shows High / Medium / Info counts and a verdict hint.
• Results are summarized with a Low / Medium / High / Critical verdict plus file/line evidence.

Cleaning up test agents

/paseo-agent-cleanup
First show a dry-run of inactive agents with disposable/test markers and test workspace candidates.
Preserve ordinary idle sessions and do not archive agents outside the IDs or patterns I specify.
Do not touch running agents, and only show workspace candidates before approval.

Core principles:

• The default run is a dry-run and does not archive ordinary idle agents based on status alone.
• --auto is allowed only for inactive agents with an explicit ID, a user-specified pattern, or a clear disposable/test/validation marker.
• It never deletes, only archives, and always protects active agents.
• Workspace archive proceeds only with user approval and an explicit ID.
• It confirms the target ID and status in the archive response and re-queries that the agent disappeared from the active list.

---

/paseo-share

Publishes small documents, code, PDFs, and images as immutable artifacts to a dedicated private Git repository. Without GitHub Actions or a background daemon, each computer only runs local git fetch, rebase, commit, and push.

• onboard: check GitHub authentication, then create <login>/paseo_share private repo or safely reuse it
• setup: configure the per-machine name and the same remote repository
• publish: return a share ID, preview URL, and download URL
• list / latest: view another computer's artifacts after remote sync
• fetch: verify and copy a share ID or latest to the current computer

Artifacts are stored under artifacts/<machine>/<year>/<month>/<artifact-id>/. Only ordinary documents, code, PDFs, and images at 50 MiB or less are allowed, and secret-looking file names and representative text token patterns are blocked. Fetching checks the metadata schema, path containment, symlinks, file size, and SHA-256, and safely handles Windows CRLF and macOS/Linux LF conversion.

The supported model is one user trusting their own private repository across their own devices. Multi-tenant exchange where untrusted multiple users share one repository is not supported. Sensitive information inside binaries is not inspected, and Git deletion still leaves history, so sensitive files must not be published.

Auto-onboarding aborts without attempting to change visibility, rename, delete, or overwrite when a same-name repository is public or is not a Paseo Share structure.

---

/paseo-orchestration comparison mode

This skill gives the same problem independently to several agents, then a separate judge compares them and picks the final answer.

Mode	Purpose
Debate	One model argues for, another against, and a judge synthesizes
Competing Plans	Compare several design/modification plans
Competing Implementations	Compare implementation candidates in isolated workspaces

Key rules:

• First read the notes from list_profiles and materialize the profile into create_agent arguments. If no profile fits, confirm with list_providers, inspect_provider, and, if needed, list_models.
• Analysis-only work may use the same workspace.
• A tournament that modifies files uses a separate workspace per participant.
• The judge preferably uses a different provider from the participants.
• Results are summarized as winner, runner-up, best merged plan, and risks.

---

/paseo-project

Selects project summary and environment setup from a single entry point.

• Summary mode: read-only summary of the requested repo overview, commands, and handoff only.
• Setup mode: performs only the requested initial setup, environment change, or context generation.
• Reuses information already confirmed and does not re-investigate the same documents in full.
• Queries version/daemon/workspace only when the Paseo runtime is relevant. If a --cwd lookup fails due to a path alias, it uses the existing workspace ID.
• If you asked only for a summary with no install/create work, it creates no files.

The procedure for the mode you need is in the skill and its linked references.

---

/paseo-spyware-check

Before installing a GitHub URL or local repo, statically inspects for spyware, malicious install scripts, secret theft, remote code execution, and supply-chain risk signals.

What it checks:

• preinstall, install, postinstall, prepare in package.json
• remote download/execute, hidden processes, and persistence in shell/PowerShell scripts
• access to .ssh, .aws, .npmrc, browser profiles, keychain, and API tokens
• obfuscation/execution patterns such as eval, Function, base64, EncodedCommand, child_process
• GitHub Actions risks: pull_request_target, unpinned actions, secret exposure
• optional scanner results: gitleaks, trufflehog, semgrep, osv-scanner, trivy, shellcheck, PSScriptAnalyzer, yara
• the helper report severity classes: High, Medium, Info
• the helper report Finding Summary count and verdict hint
• self-reference patterns inside scanner docs/scripts

Key rules:

• It does not execute the code of the repo under inspection.
• It does not run dependency install/build/test.
• It does not print sensitive values; it redacts around file/line/key names.
• Even a clean result is phrased as "no high-risk signals in the static scan", not "absolutely safe".
• Self-references caught in scanner documentation or regex descriptions are not hidden but classified as Info.

Third-party scanner note:

• /paseo-spyware-check calls external open-source scanner CLIs installed locally.
• Paseobility does not bundle or redistribute those scanners' binaries or rule sets.

Bundled helpers:

# Cross-platform JSON security receipt
python skills/paseo-spyware-check/scripts/spyware-check.py https://github.com/owner/repo --json

# macOS/Linux
skills/paseo-spyware-check/scripts/spyware-check.sh --target https://github.com/owner/repo
skills/paseo-spyware-check/scripts/install-scanners.sh --dry-run

# Windows PowerShell
.\skills\paseo-spyware-check\scripts\spyware-check.ps1 -Target https://github.com/owner/repo
.\skills\paseo-spyware-check\scripts\install-scanners.ps1 -DryRun

---

/paseo-agent-cleanup

Safely cleans up Paseo's inactive agents and test workspaces.

What it checks:

• the agent list from paseo ls --json
• the workspace list from paseo workspace ls --json
• inactive agents limited by an explicit ID, a user-specified --pattern, or a disposable/test/validation marker
• ordinary idle agents to preserve and active agents that are protected
• explicitly specified agent/workspace IDs

Key rules:

• The default run is a dry-run and does not archive ordinary idle agents based on status alone.
• --auto archives only inactive agents with an explicit ID, a user-specified pattern, or a clear disposable/test/validation marker.
• Agents in running, initializing, working, active, starting, queued, pending, busy, executing, or in-progress states are not archived.
• It does not use delete/stop/kill/restart or history/timeline/resume verification.
• After archiving it verifies the Paseo CLI JSON acknowledgement and removal from the active list.
• Workspace archive runs only after an explicit ID and --archive --yes or clear user approval.

Bundled helper:

node skills/paseo-agent-cleanup/scripts/agent-cleanup.js
node skills/paseo-agent-cleanup/scripts/agent-cleanup.js --dry-run --pattern 'cleanup-validation|fixture'
node skills/paseo-agent-cleanup/scripts/agent-cleanup.js --auto --agent <agent-id>
node skills/paseo-agent-cleanup/scripts/agent-cleanup.js --auto --pattern 'cleanup-validation|fixture'
node skills/paseo-agent-cleanup/scripts/agent-cleanup.js --workspace <workspace-id> --archive --yes

---

/paseo-browser

Drives a web page with CloakBrowser (stealth Chromium via the cloakbrowser
package's Playwright API) as the default backend — executable, not just advisory.

To do	Call
Read a page	page.goto(url) -> page.locator('body').ariaSnapshot()
Click a button	page.getByRole('button', { name }).click()
Fill a form	page.locator('#id').fill('...')
Select a dropdown	page.selectOption('#id', 'value')
Verify a screen	page.screenshot({ path }) + fresh ariaSnapshot()
Check responsiveness	page.setViewportSize({ width, height }) -> screenshot
Debug	page.on('console'/'requestfailed'/'response', ...)

The skill requires a Paseobility-owned runtime (below) and a signed
(Ed25519 + SHA-256 verified) stealth-Chromium binary in its runtime cache.

Key rules:

• Act with locators/roles, then re-read state (fresh snapshot) after navigation or a DOM change; a click alone is not proof.
• Default is headless and isolated; sessions never share the user's Chrome/Paseo profile, cookies, login, or refs. Close the browser in finally so failed runs do not leak it.
• Prefer the keyless free binary (macOS: Chromium 145; Linux/Windows: Chromium 146). The latest keyed build (Chromium 154 on all supported platforms) is free for one concurrent session with a GitHub-issued key, or paid for scaling — a paid plan is not mandatory to use a current build. Never run cloakbrowser login or store a license key; the wrapper's latest release is not the latest browser binary.
• Do not add a confirmation prompt for reversible steps the current browser task already authorizes; do verify payment/submission/account changes.
• Do not read sensitive information such as cookies, tokens, or localStorage with evaluate.

Setup — Paseobility-owned CloakBrowser runtime (no global install)

RUNTIME="${PASEOBILITY_CLOAK_RUNTIME:-$HOME/.local/share/paseobility/browser/cloakbrowser}"
mkdir -p "$RUNTIME" && cd "$RUNTIME" && npm init -y >/dev/null
npm install --save-exact --no-audit --no-fund [email protected] [email protected]
export PASEOBILITY_CLOAK_RUNTIME="$RUNTIME"
export CLOAKBROWSER_CACHE_DIR="${CLOAKBROWSER_CACHE_DIR:-$RUNTIME/cache}"  # runtime-scoped binary cache
export CLOAKBROWSER_AUTO_UPDATE=false            # reproducible pin

The package supports ESM and CommonJS. Recipes use dynamic import() (Node ^20.19.0 or >=22.12.0; verified on Node 22.23.2). See skills/paseo-browser/references/cloakbrowser.md.

Skill copying stays docs-only: the convenience installers do not install
Node packages or download the browser binary, including for a temp --target-home.

Optional compatibility backends (not the default)

The Paseo browser_* tools (live Paseo workspace tabs; needs a connected desktop
host; operational guidance in
skills/paseo-browser/references/paseo.md)
and the pinned Microsoft Playwright CLI remain available when the user opts in.
They are never silently substituted when CloakBrowser fails. Playwright CLI
recipe (@playwright/[email protected] via npx, Node >=18, no global install),
isolated headless runs with tracing/PDF/console/network. Replace S with a
task-unique value and keep -s=... quoted:

TASK_DIR="$(mktemp -d "${TMPDIR:-/tmp}/paseo-pw-XXXXXX")"
cd "$TASK_DIR"
export PLAYWRIGHT_BROWSERS_PATH="$TASK_DIR/browsers"
S='task-browser-demo-20260930'
# one-time for this task dir if no usable system Chrome:
npx --yes --package @playwright/[email protected] playwright-cli install-browser chromium
npx --yes --package @playwright/[email protected] playwright-cli "-s=$S" open 'https://example.com/' --browser chromium
npx --yes --package @playwright/[email protected] playwright-cli "-s=$S" snapshot
npx --yes --package @playwright/[email protected] playwright-cli "-s=$S" tracing-start
# ... bounded step to evidence ...
npx --yes --package @playwright/[email protected] playwright-cli "-s=$S" tracing-stop
npx --yes --package @playwright/[email protected] playwright-cli "-s=$S" close

This runs from the task dir so default .playwright-cli/ artifacts stay there.
On Windows, create the task dir and session with PowerShell ($TaskDir = Join-Path $env:TEMP ..., then Set-Location -LiteralPath $TaskDir,
$env:PLAYWRIGHT_BROWSERS_PATH, $Session) and pass "-s=$Session"; the npx
lines are otherwise the same. Playwright refs (e12)
are a separate namespace from Paseo refs (@e12); pick one backend per flow. The
mode never attaches to the user's existing Chrome profile, cookies, or tabs,
closes only its own named session (never close-all/kill-all), and adds no
installer, global-package, or MCP side effects. CloakBrowser recipes are in
skills/paseo-browser/references/cloakbrowser.md;
Playwright CLI recipes are in
skills/paseo-browser/references/playwright.md.

---

/paseo-cua

Drives a native desktop GUI through the trycua Cua Driver (cua-driver CLI or MCP), explicitly only. Preview / limited validation — see the Cua platform validation status.

• Preconditions: cua-driver --version must be present. If missing, report the prerequisite and stop; the skill installs nothing.
• Read the installed contract first: cua-driver list-tools, cua-driver describe <tool>.
• cua-driver mcp-config --client opencode only prints a config snippet; it does not register a server.
• Core loop: inspect -> snapshot -> one snapshot-bound action -> verify.
• Escalate to delivery_mode: "foreground" only after a verified background no-op and only within the user's authorization.

---

/paseo-orchestration

A pattern collection where one agent creates several agents, splits roles, and synthesizes results.

Pattern	Purpose
Fan-out	Give independent work to several agents at once
Task DAG	Step 1 -> Step 2 -> Step 3 sequential execution
Hybrid DAG	Mixed flow such as parallel work then synthesis then review
Decision Gate	A review agent decides pass/block before the next step
Coordinator Loop	Periodically check long work with a heartbeat
Blocking Ask/Reply	Proceed after receiving another agent's answer
Escalation	Clearly report permissions, ambiguity, and hard failures to the user

Recommended composition:

Codex       -> implementation / refactoring / test writing
Claude      -> review / UX copy / risk check
Coordinator -> decomposition / progress management / final synthesis

Safety rules:

• If several workers may modify the same file, create a separate workspace.
• Read the notes from list_profiles first, then materialize the chosen profile into provider/settings. Do not reuse stale provider strings verbatim.
• Set maxRuns or expiresIn on heartbeats and schedules.
• Retry transient failures such as network timeouts at most once.
• Escalate insufficient permissions, ambiguous requirements, and destructive operations to the user instead of guessing.

---

Optional extra guidance

In Paseo 0.6, the agent profiles configured in the app are the basis for provider/model/mode/thinking/feature selection. The legacy file below can be used to pass extra user guidance, not as a provider source.

~/.paseo/orchestration-preferences.json example:

{
  "preferences": [
    "Write task instructions as a self-contained briefing.",
    "Prefer a different provider for review agents than the implementing agent.",
    "Do not silently bypass hard failures; report them to the user."
  ]
}

---

Repository layout

skills/
├── paseo-agent-cleanup/       # SKILL.md + CLI helper/tests
├── paseo-browser/            # SKILL.md + references (cloakbrowser, paseo, playwright, screenshots)
├── paseo-cua/                # SKILL.md + explicit-only policy
│   └── references/           # setup.md, workflow.md, recovery.md
├── paseo-design/             # SKILL.md + references, templates, scripts/design-check.mjs, vendored guides
├── paseo-god/                # SKILL.md + explicit-only policy, procedures, run helper/tests
├── paseo-multi/              # SKILL.md + host adapter/tests + references
├── paseo-orchestration/      # SKILL.md + explicit-only policy
│   └── references/           # coordination.md, tournament.md
├── paseo-project/            # SKILL.md
│   └── references/           # brief.md, setup.md
├── paseo-share/              # SKILL.md + CLI helper/tests
└── paseo-spyware-check/      # SKILL.md + scanners/tests
scripts/                     # installers, doctor (.sh/.ps1/.py), context, cua-driver helpers
tests/                       # test_doctor.py, test_e2e_assets.py, test_cua_driver_runtime.py/.ps1, test_skill_migration.py
e2e/                         # browser fixture, fixture server, runbook, report template
.github/workflows/ci.yml     # macOS + Windows offline suites
docs/compatibility-0.10.2.md
docs/compatibility-0.9.0-beta.2.md   # historical (0.9.0-beta.2)
docs/cua-platform-validation.md
AGENTS.md
CLAUDE.md
VERSION
paseobility.json

---

License

MIT License.

---