codex-workflow-recommender

Pass

Audited by Gen Agent Trust Hub on Jul 11, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes standard system tools and shell commands (find, grep, git, ls) and Codex CLI commands to discover project structure and verify installation health. These operations are limited to informational discovery and environment profiling as defined in the permitted tool scope.
  • [PROMPT_INJECTION]: The skill is exposed to indirect prompt injection by reading untrusted data from the analyzed repository (e.g., AGENTS.md, package.json, and source documentation). This risk is mitigated by the skill's read-only design and the absence of tools for network exfiltration or persistent file modification.
  • [CREDENTIALS_UNSAFE]: The skill inspects local configuration and environment files (e.g., ~/.codex/config.toml, .env) to identify current settings. It explicitly includes safety instructions for the agent to avoid recommending secret exposure and to prioritize secure credential management via environment variables.
  • [SAFE]: The skill adheres to a least-privilege, read-only posture. It explicitly forbids the agent from performing installation, removal, or reconfiguration tasks, ensuring that all workflow improvements are presented only as recommendations for user review.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 11, 2026, 04:49 PM
Security Audit — agent-trust-hub — codex-workflow-recommender