Skillget

thegitai-cli

mcp
Guvenlik Denetimi
Basarisiz
Health Uyari
  • License — License: NOASSERTION
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Low visibility — Only 6 GitHub stars
Code Basarisiz
  • process.env — Environment variable access in packages/source-visible-customer-cli/bin/ai.ts
  • fs.rmSync — Destructive file system operation in packages/source-visible-customer-cli/bin/browser-host.ts
  • process.env — Environment variable access in packages/source-visible-customer-cli/bin/browser-host.ts
  • spawnSync — Synchronous process spawning in packages/source-visible-customer-cli/scripts/copy-tui-binary.mjs
  • process.env — Environment variable access in packages/source-visible-customer-cli/scripts/copy-tui-binary.mjs
  • process.env — Environment variable access in packages/source-visible-customer-cli/src/api/auth.ts
  • execSync — Synchronous shell command execution in packages/source-visible-customer-cli/src/api/browser-login.ts
  • os.homedir — User home directory access in packages/source-visible-customer-cli/src/api/browser-login.ts
Permissions Gecti
  • Permissions — No dangerous permissions requested

Bu listing icin henuz AI raporu yok.

SUMMARY

AI coding agent for your terminal: understands your repo, edits across files, runs and verifies changes with your approval. Source-visible CLI client.

README.md

TheGitAI - AI coding agent for your terminal

TheGitAI is an AI coding agent for terminal-based engineering work. Give it the
outcome you want: it investigates the repository, changes code across files,
runs the real project, verifies the result, and explains what is ready to
review.

TheGitAI finding and fixing a failing test in a real repository

A real session: a failing test, traced to its cause, fixed, and verified.

Documentation ·
Features ·
Website

Quick start

npm i -g @thegitai/cli
cd your-project && ai

ai signs you in on the way if you aren't already. Node.js 22 or newer is
required, and Node.js 24 is recommended. If ai is already taken on your
machine, thegitai starts the same CLI.

See Installation and
Authentication.

Ask for the finished outcome

You don't need to break a task into a chain of tiny prompts. Start with the
result and the evidence you expect:

Trace the failing signup test, fix the underlying validation, run the focused
suite, check the final diff for regressions, and update the signup guide.

What it can do

  • Understand - search the repository, read the relevant files and related
    tests, and gather current web context.
  • Change - create, patch, replace, and delete files across an
    implementation, its tests, configuration, and documentation.
  • Run - use the shell, diagnostics, scratch scripts, and managed background
    processes in the real project environment.
  • Verify - run tests, builds, linters, typecheckers, and final-diff checks
    as evidence for the handoff.
  • Work beyond source - read PDFs, spreadsheets, and Word documents; edit
    visible DOCX text; inspect screenshots; and fetch public web pages.
  • Stay controllable - ask for approval, work read-only in Plan mode, undo
    assistant edits, restore checkpoints, and resume repo-scoped sessions.

The full manual is What It Can Do.

Highlights

  • Task list and parallel work - a visible task list for multi-step work, up to six read-only lookups at
    once, and a queued follow-up message while the agent is still working.
  • Background jobs - dev servers, watchers, and builds keep running with a live tail of their
    output. /jobs browses and stops them.
  • Documents, images and web - PDF, XLSX, and DOCX reading, DOCX text edits, pasted screenshots, and
    public web search and page fetches.
  • MCP servers - connect local stdio or remote HTTP tools with ai mcp add. /mcp shows
    their status and tools.
  • Browser control - with the TheGitAI Chrome extension, the agent works in the Chrome you
    already have open, signed in as you: cloud consoles, dashboards, and admin
    panels with no API.
  • Saved sessions - sessions are saved locally per repository. Resume one with /resume or
    ai --session <id|name>.

Modes and safety

Mode Behavior
Default · Ask for Approval Asks before shell commands, file edits, browser actions, and MCP tool calls.
Bypass all permissions Approves those actions for the rest of the session. Start with ai -y. Blocked commands stay blocked.
Plan Read-only. The agent inspects, asks questions, and plans, but does not edit files or run project code.

File and shell operations are confined to the repository. Assistant edits can
be undone and checkpoints restored. When an approved sudo command needs a
password, the prompt stays masked and local and the password is never sent to
the model or the server.

See Sessions & Modes,
Approvals & Guardrails,
and Trust & Privacy.

Everyday commands

ai                       start a session in the current repo, signing in if needed
ai "<request>"           start a session with <request> as the first message
ai -y                    start in Bypass all permissions mode
ai --list-sessions       list saved sessions for this repo
ai --session <id|name>   resume a saved session
ai mcp add ...           connect an MCP server
ai --help                show commands, modes, keys, and options

Inside a session: /model picks a model and its reasoning effort, /jobs
manages background jobs, /mcp checks MCP connections, /resume reopens a
saved session, /new starts a fresh conversation, and /help shows the rest.

The complete list is in the CLI Reference.

Documentation

About this repository

This repository is source-visible so you can inspect the client that runs on
your machine: the TypeScript CLI and the Rust terminal UI, local file and shell
tools, approvals, path containment, redaction, patch application, undo, and
local session storage. Server-owned behavior is not included.

To report a vulnerability, see SECURITY.md.

License

Proprietary and source-visible. This source is published for inspection and
trust verification only; it is not open source. Copying, redistribution,
commercial use, sublicensing, derivative distribution, or use as a standalone
cloned product or hosted service is not permitted. See LICENSE.

Yorumlar (0)

Sonuc bulunamadi