fnl-committee
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied problem descriptions through the
$ARGUMENTSplaceholder and distributes them to multiple reasoning models. This creates a surface where malicious instructions in the problem description could influence the behavior of the committee members. - Ingestion points:
SKILL.md(via$ARGUMENTSin the user context). - Boundary markers: Absent for the initial user input interpolation.
- Capability inventory: The skill coordinates sub-agents that use
fnl-delegate, which has capabilities to execute commands and modify files. - Sanitization: No explicit sanitization or validation of the input is specified.
- [EXTERNAL_DOWNLOADS]: The skill provides instructions to download and install a required dependency (
fnl-delegate) using a package runner command. - Evidence:
npx skills add finolabs/finolabs-vibecode --skill fnl-delegate. - The referenced repository (
finolabs/finolabs-vibecode) belongs to the skill's author (finolabs).
Audit Metadata