The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes various local commands to set up the testing environment, including Git operations (git checkout, git diff), GitHub CLI commands (gh pr view, gh issue edit), and development server startup scripts (bin/dev, npm run dev). It also drives browser automation via the agent-browser binary.
[PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection. It ingests untrusted data from external sources including Git diffs, Pull Request metadata, and issue descriptions.
Ingestion points: The skill reads external data via git diff outputs, gh pr view output, and repository documentation such as VISION.md and PERSONAS.md (via gh issue view).
Boundary markers: The instructions do not define explicit delimiters or instructions to ignore embedded commands within the ingested data, which could lead the agent to follow instructions contained within a PR description or code comments.
Capability inventory: The agent has significant capabilities including the authority to modify source code, commit changes, execute shell commands, and interact with web browsers.
Sanitization: There is no evidence of sanitization or filtering of the external content before it is interpolated into the agent's context to determine next steps in the 'Fix Loop'.

dogfood