grill-me

Pass

Audited by Gen Agent Trust Hub on May 15, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill consists entirely of instructional text and does not execute any shell commands, access sensitive local files, or download external content.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it is designed to process user-provided design plans through multiple reviewer agents. 1. Ingestion points: User plans or designs provided during Step 1 and Step 3 of the workflow in SKILL.md. 2. Boundary markers: The skill does not define delimiters or provide instructions to isolate user input from the agents logic. 3. Capability inventory: The skill initiates parallel tool calls to reviewer personas (plan-product-hat, plan-engineering-hat, plan-design-hat) which process the ingested plan content. 4. Sanitization: No sanitization or validation is applied to the user-provided plan before it is fanned out to other agent tools.
Audit Metadata
Risk Level
SAFE
Analyzed
May 15, 2026, 09:06 AM
Security Audit — agent-trust-hub — grill-me