sde
Pass
Audited by Gen Agent Trust Hub on Nov 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes decision and state files from the .nugelo/ directory to maintain context across tasks. This ingestion of external data creates a potential surface for indirect prompt injection.\n- Ingestion points: Decisions and handoff files located in the .nugelo/ directory (referenced in SKILL.md).\n- Boundary markers: The skill includes a dedicated 'Project State Protocol' section that warns the agent to treat state files as untrusted data and not as direct instructions.\n- Capability inventory: The agent is granted capabilities for hands-on coding, including writing features, fixing bugs, and implementing APIs.\n- Sanitization: The instructions mandate that any directives found within processed state files must be surfaced to the user rather than followed by the agent.
Audit Metadata