The Agent Skills Directory

[SAFE]: The skill does not contain any patterns associated with data exfiltration, obfuscation, or unauthorized remote code execution.
[COMMAND_EXECUTION]: The skill uses gh (GitHub CLI) and git commands for managing issues and labels. These operations are restricted to project management tasks and do not involve high-risk system modifications.
[PROMPT_INJECTION]: The skill processes untrusted conversation history, creating an indirect prompt injection surface. This is mitigated by a mandatory human-in-the-loop review process:
Ingestion points: Conversation history is reviewed in Step 2 to identify feedback signals.
Boundary markers: No explicit delimiters are used to wrap the ingested session content.
Capability inventory: The skill can create GitHub issues (gh issue create), create labels (gh label create), and search issues (gh search).
Sanitization: Step 4 uses AskUserQuestion with a multi-selection interface, ensuring that no issues are created without explicit user verification and approval of the analyzed findings.

feedback-session