The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to read external data from GitHub issue and pull request comments, which could potentially contain adversarial instructions aimed at influencing the agent's behavior.
Ingestion points: External content is fetched in discovery.md and tickets.md using gh issue view <n> --comments and gh pr view <pr> --comments.
Boundary markers: The skill does not define specific delimiters or instructions for the agent to ignore potentially malicious embedded content within these comments.
Capability inventory: The skill has access to powerful tools, including git (commit, push, branch management) and gh (issue creation, pull request review/approval, and deployment management).
Sanitization: There is no explicit mechanism described to sanitize or filter the comment data before the agent processes it, though the skill mandates that a human expert must explicitly confirm actions in the current session before sensitive operations like gh pr review --approve are executed.
[COMMAND_EXECUTION]: The skill makes extensive use of shell commands through git and the GitHub CLI (gh) to manage project documentation and repository state. These commands are fundamental to the skill's primary purpose of automating product management workflows.
[PRIVILEGE_ESCALATION]: The onboarding.md file includes instructions for the user to perform administrative installation tasks, such as sudo apt install gh, to set up the necessary environment for the skill.
[EXTERNAL_DOWNLOADS]: The skill references external downloads for its own updates via npx and for system dependencies via official package managers such as Homebrew, APT, and Winget. These references target well-known services and the author's own repository.

boring-pm