signadot-plan

Pass

Audited by Gen Agent Trust Hub on May 14, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface where untrusted data from the Signadot catalog could influence agent behavior.\n
  • Ingestion points: The agent retrieves potentially untrusted metadata (names, descriptions, and code bodies) from the Signadot platform via signadot plan action list and signadot plan action get (SKILL.md).\n
  • Boundary markers: There are no instructions or patterns provided to the agent to treat this external data as untrusted or to wrap it in delimiters to prevent the obedience of embedded instructions.\n
  • Capability inventory: The agent possesses the capability to create and run plans (signadot plan create, signadot plan run), which allows it to execute logic in the target cluster environment based on its interpretation of the action catalog (SKILL.md).\n
  • Sanitization: No sanitization, validation, or escaping of the retrieved action data is specified before the agent uses it to compose or iterate on plan specifications.
Audit Metadata
Risk Level
SAFE
Analyzed
May 14, 2026, 12:19 AM
Security Audit — agent-trust-hub — signadot-plan