harness-setup
Pass
Audited by Gen Agent Trust Hub on Aug 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill performs standard repository configuration tasks, including stack detection and the generation of architectural documentation. It follows a structured, phase-based workflow that requires human review before finalizing configuration changes.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests information from local project files, such as plan.md and architecture reports, to populate instructions in CLAUDE.md and other guides. This represents a data ingestion surface, but the impact is limited to the local development environment and is overseen by the user.
- Ingestion points: plan.md, architecture-design report, tech-stack-comparison report.
- Boundary markers: Not explicitly defined in the skill text.
- Capability inventory: The skill is limited to reading and writing files within the project workspace and .ai/workspace/harness/ directory.
- Sanitization: No explicit sanitization or filtering of ingested project data is performed before interpolation.
Audit Metadata