The Agent Skills Directory

[SAFE]: The skill defines clear protocols for repository interaction using standard tools and environment variables. No malicious behaviors or unauthorized access patterns were identified.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from GitHub (issues, PRs, comments). 1. Ingestion points: Repository context is read via orchestrator_read as described in SKILL.md. 2. Boundary markers: The skill instructs the agent to ignore bot-generated words for authorization purposes (bot: true). 3. Capability inventory: The agent can execute gh api commands to post replies and call orchestrator_report as seen in SKILL.md. 4. Sanitization: The instructions recommend using gh api with field flags (-f) for replies, which is a safer pattern for data interpolation.

dsh-superagent-github