implement
Pass
Audited by Gen Agent Trust Hub on Jul 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues detected. The skill utilizes a gated workflow designed to ensure all code implementation tasks have been previously approved by the user in a local state file.
- [PROMPT_INJECTION]: The skill reads implementation tasks and project documentation from local markdown files, representing an indirect prompt injection surface where instructions in data files could influence agent behavior. This is the primary intended function of the skill and is mitigated by a mandatory state-check gate that requires manual approval of the project plan and specification before any tasks are processed.
- Ingestion points: tasks.md, plan.md, data-model.md, contracts/
- Boundary markers: Absent; the skill relies on the SpecForge approval state check
- Capability inventory: Read, Write, Edit, Glob, Grep, Bash
- Sanitization: Absent; the skill relies on the preceding manual approval process to validate the contents of the task files
Audit Metadata