pm-brainstorming

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill contains an attack surface for indirect prompt injection due to its requirement to ingest untrusted data from local project files and documentation.
  • Ingestion points: The agent is instructed in Phase 1 to read and process activeContext.md, insights.md, and any existing specifications or research related to the topic.
  • Boundary markers: No specific delimiters or instructions to ignore embedded commands (e.g., 'ignore previous instructions') are provided when processing these external documents.
  • Capability inventory: The skill's potential for misuse is low as its toolset is limited to file reading (Read, Grep, Glob) and user communication (AskUserQuestion).
  • Sanitization: The instructions do not include steps to sanitize or validate the content of the read files before they are integrated into the agent's brainstorming logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 06:37 AM
Security Audit — agent-trust-hub — pm-brainstorming