mold
Pass
Audited by Gen Agent Trust Hub on Jun 20, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements strong security controls to govern its operations. The 'two-key handshake' requires explicit user confirmation before the skill writes any specifications or issues to the project directory. Additionally, the 'coherence self-check' ensures that the agent verifies the logical consistency and grounding of a plan before attempting to conclude a session.
- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it is designed to ingest and process external content such as stack traces, error reports, and external research output. However, this risk is categorized as safe due to the following factors:
- Context: Data ingestion is fundamental to the skill's primary purpose of debugging and planning.
- Human-in-the-loop: All outputs and tool invocations involving state changes require explicit user authorization, preventing autonomous exploitation of ingested instructions.
- Capability Gating: The skill strictly delegates specialized research tasks to separate sub-agents with their own constraints.
Audit Metadata