task-to-pr
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user requirements, tickets, and issues which are external inputs that could contain instructions.
- Ingestion points: Phase 1, Step 1 reads requirements from tickets and issues within the SKILL.md file.
- Boundary markers: The skill does not define specific prompt delimiters or instructions to ignore embedded commands within the external ticket content.
- Capability inventory: Includes git command execution (commit, push), pull request creation using
gh, and execution of local test/build scripts (e.g., npm, pnpm, yarn, pytest, go test, cargo test, mvn test, gradlew). - Sanitization: The workflow relies on manual diff review (Phase 5) and self-review steps to verify changes before they are finalized.
- [DYNAMIC_EXECUTION]: The "Self-Improvement" section contains instructions for the agent to patch the skill file itself to refine the workflow.
- Pattern: The agent is encouraged to update the SKILL.md instructions if validation gates or delivery rules are found to be missing during execution.
- Context: This is a transparent self-optimization pattern for developer workflows and does not involve the execution of arbitrary code generated from untrusted external inputs.
- [SAFE]: The skill uses legitimate development tools and system commands for their intended purposes. No evidence of credential harvesting, unauthorized data exfiltration, or malicious obfuscation was found.
Audit Metadata