The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the GitHub CLI (gh) to perform repository management tasks.
The agent is instructed to use gh issue list, gh label list, gh issue search, and gh issue create to manage tickets on GitHub.
These commands are consistent with the primary purpose of the skill and target the service's official API/CLI endpoints.
[EXTERNAL_DOWNLOADS]: The documentation references external resources for technical context and development guidelines.
Links are provided to official documentation and the primary GitHub CLI repository (github.com/cli/cli). These are recognized as trusted or well-known services and do not escalate the security risk.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection via the processing of untrusted user input.
Ingestion points: User-provided failure descriptions, reproduction steps, and expected outcomes are ingested into the agent context in SKILL.md (Step 1).
Boundary markers: The agent uses prescribed Markdown templates to encapsulate user content when filing issues, although it does not utilize explicit secondary delimiters (like XML tags) for the input itself.
Capability inventory: The skill employs the gh tool for writing to external trackers and a sub-agent (subagent_type=Explore) for reading codebase data.
Sanitization: No programmatic sanitization is applied to user strings; the agent relies on its inherent instructions to clarify and refactor user feedback into domain-appropriate language before submission.

qa