planner

Pass

Audited by Gen Agent Trust Hub on Aug 27, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill interacts with external data and web content via BrowserOS, establishing an attack surface for instructions embedded in third-party data.
  • Ingestion points: External data and web content (SKILL.md).
  • Boundary markers: Instructs the agent to treat web content as hostile.
  • Capability inventory: Subprocess calls via ExecutionProxy, file-write capabilities via MemoryStore, and cognitive tools like sequentialthinking.
  • Sanitization: Mandates the redaction of secrets and PII before sharing context.
  • [COMMAND_EXECUTION]: The skill uses an ExecutionProxy to perform terminal actions such as 'npm test' to verify code integrity through TDD.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 27, 2026, 07:35 AM
Security Audit — agent-trust-hub — planner