hedge-fund-committee-eval

Pass

Audited by Gen Agent Trust Hub on Aug 20, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a local JavaScript file apply_score_caps.mjs using Node.js to enforce scoring penalties. This script processes JSON data from the workflow run to ensure the final score is mathematically consistent with defined rubric caps.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external workflow data (runJson) which may contain untrusted content from news sources (FT, WSJ, Reddit).
  • Ingestion points: Data enters the process via the runJson variable passed to the scoring script and the LLM judge.
  • Boundary markers: None explicitly defined in the prompt interpolation.
  • Capability inventory: The skill has the ability to write logs to a local CSV file (reports/hedge-fund.eval.csv) but lacks network or system-level write capabilities.
  • Sanitization: No specific sanitization or filtering of the workflow output is mentioned before evaluation.
  • [PERSISTENCE_MECHANISM]: The documentation instructs the user to modify the local git configuration (git config core.hooksPath hooks) to enable repository-specific git hooks. While this is a standard practice for development workflows to prevent regressions, it directs the environment to execute scripts located in the hooks/ directory during git operations.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 20, 2026, 02:21 PM
Security Audit — agent-trust-hub — hedge-fund-committee-eval