skills/sztadii/agent-skills/council/Gen Agent Trust Hub

council

Pass

Audited by Gen Agent Trust Hub on Nov 6, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data in the form of code diffs, commit ranges, pull requests, and proposals.
  • Ingestion points: SKILL.md workflow indicates the skill reads diffs and code directly touched by changes.
  • Boundary markers: No explicit delimiters or instructions to ignore embedded malicious prompts within the processed code are defined.
  • Capability inventory: The skill uses codex exec and claude commands to perform reviews and allows "isolated read-only probes."
  • Sanitization: No sanitization or filtering of the code content is described before passing it to sub-agents.
  • [DYNAMIC_EXECUTION]: The skill dynamically constructs shell commands to invoke review agents using variables for the model, effort levels, and output paths.
  • Evidence: The workflow template uses placeholders like codex exec -m <model> -c model_reasoning_effort=<effort> .... If a user overrides the default lineup with malicious strings, it could lead to command injection in the host environment.
  • [COMMAND_EXECUTION]: The skill's primary function involves executing system commands (codex and claude) to manage the review process.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 6, 2026, 05:35 PM