The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by design, as it ingests and processes untrusted data from GitHub issues and repository templates to guide its actions.
Ingestion points: The skill reads external content using gh issue view and accesses repository-specific files in .github/ISSUE_TEMPLATE/ and .github/PULL_REQUEST_TEMPLATE.md.
Boundary markers: No explicit delimiters or isolation mechanisms are mentioned to separate untrusted issue content from the skill's operational instructions.
Capability inventory: The skill has the authority to execute repository-defined tasks (tests, linting, builds), perform Git operations (branching, committing, pushing), and manage GitHub pull requests.
Sanitization: There is no automated sanitization of the external data; however, the skill incorporates 'HARD-GATE' instructions that require the agent to verify all work against the physical repository state and approved plans before proceeding.

implement