adversarial-plan-review-codex

Pass

Audited by Gen Agent Trust Hub on Aug 16, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSNO_CODE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The README.md file references a GitHub Gist associated with the author (dachent) to document the provenance of the adapted planning prompts. This is a static reference to the developer's resources and does not involve automated remote code execution.
  • [PROMPT_INJECTION]: The skill is designed to ingest and analyze external, untrusted plans, which creates an inherent surface for indirect prompt injection. (1) Ingestion points: The SKILL.md instructions direct the agent to read external plan steps, success criteria, and evidence catalogs. (2) Boundary markers: The instructions do not specify the use of delimiters or 'ignore' warnings for the data being processed. (3) Capability inventory: The skill includes the capability to write review results to the file system as artifacts. (4) Sanitization: The skill does not describe any specific input validation or sanitization requirements for the plan content.
  • [NO_CODE]: The skill consists of markdown instruction files, reference guides, and metadata configuration but does not ship with any Python, JavaScript, or other executable scripts.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 16, 2026, 07:58 AM
Security Audit — agent-trust-hub — adversarial-plan-review-codex