soulforge

What This Skill Does

Soulforge is an advanced autonomous coding engine designed for OpenClaw agents to execute complex software development tasks with high precision. It orchestrates end-to-end loops including planning, implementation, verification, and iterative refinement. Unlike simple script execution, Soulforge enforces a strict operational framework that emphasizes safety through worktree isolation, preventing accidental damage to your main repository checkout. It is engineered to operate in discrete phases—feature development, bug fixes, and review loops—ensuring the agent maintains context while adhering to specific, human-defined constraints.

Installation

To integrate Soulforge into your environment, run the following command in your terminal:

clawhub install openclaw/skills/skills/jamesrp13/soulforge

Use Cases

• Autonomous Feature Development: Use the feature-dev workflow to offload the end-to-end implementation of new features, from initial scaffold to test-driven completion.
• Surgical Bug Fixes: Utilize the bugfix mode for issues that require diagnostic work and localized, minimal code patches without introducing regression.
• PR Tightening: The review-loop workflow allows the agent to iteratively act on feedback, automatically refining code to meet strict quality gates or style guidelines until a PR is ready for human merge.

Example Prompts

1. "Soulforge, start a feature-dev session for PR #455. Implement the user authentication toggle. Strictly follow the specs in the linked issue and DO NOT touch the legacy CSS files."
2. "Run a bugfix loop for issue #99. Start by writing a reproduction test case for the memory leak and ensure the fix is isolated to the service layer only."
3. "Begin a review-loop for my active worktree. Focus on addressing the linter warnings in the /src/components directory and verify all unit tests pass before finishing."

Tips & Limitations

• Isolation is Key: Never attempt to run Soulforge directly on your main branch. Always utilize isolated worktrees located in your /worktrees/ directory to prevent state corruption.
• Contractual Clarity: The quality of the output is directly proportional to the clarity of your task contract. Always define explicit 'out-of-scope' items to prevent the agent from over-engineering or refactoring unrelated modules.
• Callback-Exec: Ensure you configure your --callback-exec properly. Soulforge uses this to provide real-time status updates back to your communication channels.
• Worktree Safety: The tool automatically rejects dirty worktrees, so ensure your environment is clean before triggering a run. If you encounter issues, confirm that your task description includes specific acceptance criteria.