The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the gh CLI (gh issue list, gh project item-list) and the cat command to fetch backlog data from GitHub repositories or local markdown files. These operations are essential for the skill's primary function and are triggered only upon user request.
[DATA_EXPOSURE]: The skill includes functionality to read local files via a user-defined tasks_file path. While this is intended for backlog processing, users should exercise standard caution and ensure the path points only to the intended documents.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external data from GitHub issues and local files to perform scoring and analysis. There is a theoretical risk that malicious instructions embedded within a backlog item could attempt to influence the agent's scoring logic.
Ingestion points: Data is ingested via gh issue list and cat $TASKS_FILE in SKILL.md.
Boundary markers: No specific delimiters or "ignore" instructions are explicitly defined for the ingested data in the prompt logic.
Capability inventory: The agent has access to shell execution (gh, cat, ls) as defined in the skill logic.
Sanitization: No explicit sanitization or filtering of the ingested backlog content is mentioned.

pm-prioritize