is-this-a-problem

Pass

Audited by Gen Agent Trust Hub on May 15, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted user data (questions from communication platforms like Slack or DingTalk) and incorporates it into its logic without explicit boundary markers or delimiters.
  • Ingestion points: User-provided input via the [问题] argument and the matter-specific matter.md file.
  • Boundary markers: Absent; the skill directly evaluates user text against risk patterns.
  • Capability inventory: The skill can read and write files within the ~/.claude/plugins/config/claude-for-legal/ directory and invoke other governance tools.
  • Sanitization: None identified; the text is used directly for pattern matching and response generation.
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute other specific commands such as /product-legal:matter-workspace, /ai-governance-legal:use-case-triage, and /ai-governance-legal:vendor-ai-review. This behavior is consistent with the intended purpose of orchestrating legal workflows.
Audit Metadata
Risk Level
SAFE
Analyzed
May 15, 2026, 07:06 AM
Security Audit — agent-trust-hub — is-this-a-problem