review-roadmap

Pass

Audited by Gen Agent Trust Hub on Jul 18, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The skill includes role-play instructions that direct the agent to 'read the roadmap as the agent who will materialize and execute each child would'. While intended to define the review perspective, this role-play guidance could be influenced by malicious content within the analyzed files.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes the content of untrusted local files to generate its report.
  • Ingestion points: The agent reads the contents of roadmap.md and decisions.md within the thread directory.
  • Boundary markers: There are no explicit delimiters (such as XML tags or unique markers) used to wrap the file content to distinguish it from instructions.
  • Capability inventory: The agent can write findings to the .pending-reviews/ directory using the /emit-pending-review tool.
  • Sanitization: The skill does not implement sanitization or validation of the file content before it is processed by the model.
  • [NO_CODE]: The skill consists exclusively of markdown instructions and YAML configuration files; it does not contain or execute external scripts, binaries, or compiled code.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 18, 2026, 01:01 PM
Security Audit — agent-trust-hub — review-roadmap