The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the GitHub CLI (gh) via the Bash tool to manage issues.
Evidence: Commands such as gh issue list, gh issue view, and gh issue comment are used to interact with the repository apify/apify-mcp-server.
Security Posture: The commands are scoped to the vendor's own repository. While shell interpolation of $ARGUMENTS and <response> bodies presents a minor surface for injection, the operations are restricted to non-destructive triage tasks and are subject to human review.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from GitHub issue descriptions and comments.
Ingestion points: Fetches issue data and comments using gh issue view <number> --comments in SKILL.md.
Boundary markers: The skill does not use explicit delimiters for the issue content, but the workflow clearly separates the 'Analyze' phase from the 'Action' phase.
Capability inventory: The agent can comment on, close, and edit issues.
Sanitization: No automated sanitization is performed on the incoming text.
Mitigation: The skill strictly enforces 'Step 4: Present and get approval', requiring explicit user confirmation before executing any gh issue comment or gh issue close command. This ensures the user reviews the final payload before it is sent to the GitHub API.

bug-triage