The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting external content from repository issue templates and search results. * Ingestion points: The skill reads templates from .github/ISSUE_TEMPLATE/ and existing issues via gh issue list (SKILL.md). * Boundary markers: The instructions lack explicit boundary markers to prevent the model from following instructions embedded in the external issue data. * Capability inventory: The agent has the capability to write to the local filesystem and create remote GitHub issues via the gh CLI (SKILL.md). * Sanitization: The risk is mitigated by a mandatory human-in-the-loop step, where the user must review the drafted content and provide explicit confirmation via AskUserQuestion before any external action is taken.

create-issue