plan-interview

Pass

Audited by Gen Agent Trust Hub on May 17, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes untrusted project files to drive planning and autonomous execution.
  • Ingestion points: The agent is instructed to read README.md, AGENTS.md, and other codebase files to understand architecture and conventions during the exploration phase (SKILL.md).
  • Boundary markers: The instructions do not define delimiters or markers to isolate the content of these files from the agent's instructions, which could allow embedded malicious prompts to influence behavior.
  • Capability inventory: The skill has the capability to write files (planning documents in docs/plans/) and is explicitly instructed to 'Auto-start implementation immediately' once a plan is approved, which involves making direct changes to the codebase (SKILL.md).
  • Sanitization: The process lacks validation or sanitization of ingested content before it is used to formulate implementation steps.
  • [EXTERNAL_DOWNLOADS]: The skill can be triggered to fetch content from external network locations provided by the user.
  • Evidence: During the 'Knowledge Audit', the agent is instructed to load user-provided references, such as documentation URLs or API specifications, if they are needed to fill knowledge gaps before planning (SKILL.md).
Audit Metadata
Risk Level
SAFE
Analyzed
May 17, 2026, 07:57 AM
Security Audit — agent-trust-hub — plan-interview