task-to-pr
Pass
Audited by Gen Agent Trust Hub on Jul 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted external data from repository files and user-provided requirements, creating an attack surface for indirect prompt injection.\n
- Ingestion points: Phase 1 reads README, documentation files, contributing guidelines, and linked issue/ticket content.\n
- Boundary markers: No explicit delimiters or boundary instructions are defined to separate ingested data from agent instructions.\n
- Capability inventory: Subprocess calls (git, build/test tools), file-write operations, and network push operations.\n
- Sanitization: None; however, the skill includes explicit instructions for the agent to avoid destructive actions and to ask for clarification in ambiguous or risky scenarios.\n- [COMMAND_EXECUTION]: Executes standard development and build tools (e.g., git, npm, pnpm, yarn, pytest, cargo, mvn, gradle) for repository management, code validation, and pull request creation.\n- [DATA_EXFILTRATION]: Uses git push and gh pr create to transmit code changes and task metadata to the project's configured remote repository.
Audit Metadata