equilio-map
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [NO_CODE]: The skill consists entirely of Markdown documentation and instructions. It does not include executable scripts, binaries, or automated environment modifications, which significantly reduces the risk of direct malicious activity.\n- [SAFE]: The skill follows established development principles and promotes security-conscious behavior, such as explicitly stating that agent handoffs do not grant additional permissions. The instructions prioritize technical precision and the reconciliation of facts against evidence.\n- [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves processing external data, such as issue descriptions and UI content, which constitutes an inherent attack surface. This is mitigated by the structured 'work-artifact' contract and the requirement for human verification of the resulting system model.\n
- Ingestion points: External issue descriptions (Original Request) and web browser content used for UI behavior observation (SKILL.md).\n
- Boundary markers: The 'work-artifact' contract uses specific Markdown headers (Problem, Solution, etc.) and '---' delimiters to isolate the original request (work-artifact.md).\n
- Capability inventory: Reading files, writing local Markdown artifacts, and using a browser tool for observation (SKILL.md, work-artifact.md).\n
- Sanitization: The methodology relies on structural separation within the artifact and emphasizes that prior summaries are orientation aids, not proof of code or verification state (agent-handoffs.md).
Audit Metadata