Your Claude Engineer
https://github.com/coleam00/your-claude-engineer- Category
- AI Agents
- Rank
- No. 1493Tools index
Previous survey · No. 1500 ·
- Pricing
- Open Source
- Type
- AGENT
- Builder
- coleam00
- GitHub
- 140 stars
- Date
About
Agent harness with tools for Slack, GitHub, and Linear — turn Claude into your own end-to-end software engineer.
What it does
A Python command-line runner breaks a software project into tracked issues, then cycles fresh Claude sessions through implementation, browser checks, version control, and status updates. An orchestrator assigns focused agents to coding, Linear, GitHub, and Slack work while project state persists between sessions.
Why it's ranked here
The design connects planning, implementation, verification, and team communication into one coherent loop. Fresh clients between iterations address context exhaustion, while isolated project repositories and scoped agent roles reduce accidental overlap. The substantial setup and service dependencies keep it from being a general-purpose coding assistant.
What's good
Each specialist receives only the tools needed for its role, and model choice is configurable per agent. Generated projects get separate Git repositories. Playwright supports browser-level UI checks. The security design combines sandboxing, project-scoped file permissions, a shell-command allowlist, and pre-execution validation for sensitive commands.
Tradeoffs
Linear and an Arcade gateway are required, with separate credentials and OAuth authorization. GitHub and Slack are optional, but Slack posting needs an existing channel and GitHub automation needs an existing repository. Native Windows is unsupported. The default loop can run without an iteration limit, and first-time initialization may take over ten minutes.
How to use it well
Choose it for greenfield or extended feature work where Linear already drives execution and automated pull requests or Slack updates justify the setup. Start with a capped iteration count, review the generated repository, and expand command permissions cautiously. It does not replace repository creation, Slack channel administration, or native Windows tooling.
Technical notes+
autonomous_agent_demo.py exposes the CLI and resolves project directories and model aliases. agent.py recreates a ClaudeSDKClient for each iteration, retries failed sessions, and stops when the orchestrator emits PROJECT_COMPLETE:. client.py configures built-in tools, Playwright over stdio, Arcade over HTTP, agent definitions, sandbox settings, and a PreToolUse Bash hook. agents/definitions.py assigns separate prompts, tools, and configurable models to Linear, coding, GitHub, and Slack agents. progress.py caches Linear initialization in .linear_project.json. security.py implements command allowlisting plus extra checks for deletion, process termination, permission changes, and init scripts. test_security.py exercises command extraction and security decisions.
Observed
- Primary language
- Python
- Interface
- Command-line application
- Installation surface
- Claude Code installs through npm; Python dependencies install from a pinned requirements file
- External interface
- Uses MCP servers for Arcade over HTTP and Playwright over stdio
- Platform support
- macOS and Linux are supported; Windows requires WSL or a Linux virtual machine
- Testing structure
- Repository includes a standalone Python security test script
- Project isolation
- Each generated project is created in its own directory with a separate Git repository
Read from README.md, requirements.txt, agent.py, client.py, prompts.py, progress.py, security.py, arcade_config.py, test_security.py, authorize_arcade.py, autonomous_agent_demo.py, agents/__init__.py, agents/definitions.py, agents/orchestrator.py.
What it can do
Create and manage GitHub repositories
Repository specifications and commands → GitHub repository with configured settings
Generate and commit code to GitHub
Code requirements and specifications → Code files committed to GitHub repository
Create and update Linear issues
Issue descriptions and project requirements → Linear tickets with status updates
Send automated Slack notifications
Project updates and status messages → Slack messages to specified channels or users
Review and analyze code repositories
GitHub repository URLs or code files → Code analysis reports and suggestions
Execute end-to-end software development workflows
High-level project requirements → Complete software solutions with documentation
Tags
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.