meta-investment-committee-red-team
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external investment plans, financial models, and evidence registers. This introduces a surface for indirect prompt injection if the ingested documents contain malicious instructions. However, the skill explicitly mandates a read-only workflow and follows a rigid committee framework for validation, which limits the potential impact of adversarial content within the processed data.
- [COMMAND_EXECUTION]: The instructions reference a local script,
../../../tools/investment-committee/simulate_committee.py, for performing deterministic manifest validation and scoring. This script is part of the skill's internal toolset and is used for its primary functional purpose without indicators of malicious intent.
Audit Metadata