The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses local shell commands for repository management and code validation, including git status, git diff, ruff check, and python manage.py. These commands are executed to assess the current state and quality of pull requests.
[EXTERNAL_DOWNLOADS]: The skill includes instructions to install external helper tools such as 'deslop' and 'no-mistakes' using a local script scripts/install-external-skills.py. These external skills are part of the broader ecosystem used for code quality and safety gating.
[PROMPT_INJECTION]: The skill ingests untrusted data from pull request diffs, descriptions, and comments to determine the intent of the work. This creates an indirect prompt injection surface where malicious content in a PR could influence the agent's behavior. 1. Ingestion points: PR descriptions and diffs are analyzed in SKILL.md and references/triage-protocol.md. 2. Boundary markers: No instructions are provided to use delimiters or ignore embedded instructions within the PR content. 3. Capability inventory: The agent can execute system commands and run project tests as part of its validation workflow. 4. Sanitization: The skill does not mention any sanitization or escaping of the ingested PR content.
[REMOTE_CODE_EXECUTION]: During the validation step, the skill instructs the agent to run the project's test suite, such as pytest, on the code provided in the pull request. This involves executing untrusted code from an external author in the agent's local environment to confirm bug fixes or feature functionality.

prs