brainstorming

Pass

Audited by Gen Agent Trust Hub on Apr 11, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill employs authoritative instructions and behavioral constraints to override the agent's default operational mode, using terms like 'Non-Negotiable', 'Mandatory', and 'Hard Gate' to enforce a facilitator role.
  • [PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection through its data processing workflow.
  • Ingestion points: The process explicitly requires the agent to review existing project files, documentation, and plans in the 'Understand the Current Context' step.
  • Boundary markers: No explicit delimiters or instructions are provided to the agent to treat external file content strictly as data rather than instructions.
  • Capability inventory: The skill restricts the agent to structured dialogue and writing markdown documentation; it does not request capabilities for command execution, file-system modification, or network access.
  • Sanitization: The instructions do not define methods for validating or filtering the content of ingested project files or user-provided ideas.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 11, 2026, 06:19 PM
Security Audit — agent-trust-hub — brainstorming