devils-advocate

Pass

Audited by Gen Agent Trust Hub on Sep 24, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill consists entirely of behavioral instructions and research references designed to improve decision quality through critical analysis. No executable code, network operations, or sensitive file access patterns are present.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process user-provided decisions and facts. While it lacks external tools or side-effect capabilities (such as network access or file system writes) that would allow for exploitation, it relies on user-supplied context to generate its objections.
  • Ingestion points: User input defining a 'decision, plan, belief, hire, pricing move, strategy, or piece of work' (SKILL.md, Workflow step 2).
  • Boundary markers: The prompt does not explicitly define delimiters for user input, relying on the model's standard conversational context.
  • Capability inventory: The skill is limited to text generation and does not request tool access or execute shell commands.
  • Sanitization: None present; the skill treats user input as the ground truth for its analytical exercise.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 24, 2026, 12:50 AM
Security Audit — agent-trust-hub — devils-advocate