fcc-refine-idea
Pass
Audited by Gen Agent Trust Hub on Aug 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user interview responses and explores local codebase or vault files to synthesize a project brief.\n
- Ingestion points: The agent processes user input and local file content (SKILL.md).\n
- Boundary markers: There are no explicit delimiters used to separate user-provided data from system instructions.\n
- Capability inventory: The agent is authorized to write markdown files to the
docs/briefs/directory (SKILL.md).\n - Sanitization: No specific filtering or escaping is applied to the ingested data before it is included in the output brief.\n- [DATA_EXFILTRATION]: The skill is designed to read local project information and generate briefs within a specified local directory. These actions represent standard documentation workflows and do not involve sending data to external or untrusted destinations.\n- [EXTERNAL_DOWNLOADS]: The skill contains references to official freeCodeCamp domains and its GitHub organization. These resources are well-known and trusted, and their inclusion does not pose a security risk.
Audit Metadata