vendor-ai-review

Pass

Audited by Gen Agent Trust Hub on May 15, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection (Category 8) because it ingests untrusted data (AI vendor contracts) and processes it with high-privilege context from local files.
  • Ingestion points: Untrusted data enters the agent context through the pasted contract terms.
  • Boundary markers: The instructions lack delimiters or explicit headers to isolate the untrusted contract text from the skill's logic.
  • Capability inventory: The skill reads sensitive local configuration files (~/.claude/plugins/config/claude-for-legal/ai-governance-legal/CLAUDE.md and matter.md) that contain risk thresholds and legal positions.
  • Sanitization: There is no evidence of input validation or sanitization.
  • [NO_CODE]: The skill contains only instructional markdown and does not include any Python, Node.js, or shell scripts.
Audit Metadata
Risk Level
SAFE
Analyzed
May 15, 2026, 07:07 AM
Security Audit — agent-trust-hub — vendor-ai-review