The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data retrieved from GitHub issue and pull request content, which is then used by the agent to generate a summary for the user.
Ingestion points: scripts/board.py fetches titles, bodies, and comments from issues and PRs via the gh command-line tool (lines 35, 37, and 44).
Boundary markers: The skill's instructions do not provide delimiters or instructions for the agent to distinguish between its own system prompt and potentially adversarial content within the ingested data.
Capability inventory: The skill executes scripts/board.py, which invokes the gh utility via subprocess.run and reads local repository files.
Sanitization: The Python script truncates comment bodies and PR decisions to fixed lengths (lines 103 and 127), but it does not sanitize the text for prompt injection patterns.
[COMMAND_EXECUTION]: The script scripts/board.py invokes the gh (GitHub CLI) utility using subprocess.run to list repository data. This is an expected behavior for the skill's stated purpose of reporting on project board status and uses standard GitHub CLI functionality.

board