state-modeling

Pass

Audited by Gen Agent Trust Hub on Jul 4, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection attack surface by processing external data to drive its workflow.
  • Ingestion points: The skill reads 'Target UI flow' (provided by the user) and the 'REPO_PROFILE.json' file to understand the project structure.
  • Boundary markers: No explicit delimiters or warnings are used to separate untrusted inputs from the agent's instructions.
  • Capability inventory: The skill operates in 'write' mode and is instructed to 'validate with repo commands', implying access to the filesystem and shell.
  • Sanitization: No input validation or escaping mechanisms are described.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 4, 2026, 08:23 PM
Security Audit — agent-trust-hub — state-modeling