deep-architecture-research
Pass
Audited by Gen Agent Trust Hub on Aug 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a robust 'confirmation gate' strategy, requiring the agent to present a research map (framing brief) and obtain user consent before performing any substantive research tasks. This prevents unauthorized tool usage or unintended data collection.
- [SAFE]: The 'Evidence Protocol' promotes high-integrity research by prioritizing actual source code and official documentation over secondary community signals. It explicitly instructs the agent to separate verified facts from historical data or roadmap plans.
- [SAFE]: The skill encourages the use of local clones for source code analysis rather than browsing individual web pages, which is a best practice for understanding complex system architectures and call paths without triggering excessive external requests.
- [SAFE]: Data handling is restricted to creating durable Markdown artifacts within the project's research location, with instructions to keep scratch notes out of shared paths, supporting good data hygiene.
- [SAFE]: No evidence of prompt injection, obfuscation, or unauthorized network activity was found. The instructions focus entirely on improving the quality and traceability of technical research.
Audit Metadata