The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the GitHub CLI (gh) to manage issues and labels within the chrislacey89/skills repository. It performs read operations to search for existing issues and write operations to file new proposals and create tracking labels.
[EXTERNAL_DOWNLOADS]: The skill interacts with GitHub services via the CLI and utilizes a local /library command to query and load software engineering principles from a local index.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface (Category 8) because it ingests data describing 'field incidents' from external projects and repositories. This potential risk is addressed through structured reasoning and manual verification.
Ingestion points: In SKILL.md (Phase 1), the agent captures incident details from arbitrary repositories, PRs, and execution logs.
Boundary markers: The skill does not explicitly use delimiters or instructions to disregard potential commands embedded within incident data.
Capability inventory: The skill has the capability to create GitHub issues and labels (Phase 5) and can perform file-system modifications in the target repository if the user approves the follow-on implementation (Phase 7).
Sanitization: There is no automated sanitization or filtering performed on the external incident descriptions before they are processed by the reasoning agents.

improve-pipeline