Bridges AI agents with the Codex CLI while preserving session continuity. Creates real, resumable Codex CLI sessions via MCP calls that survive server restarts. Provides three primary tools: general conversation (codex_chat), structured planning with critique (codex_plan), and code review (codex_review). Features automatic attribution distinguishing agent requests from human commands, per-request model and reasoning effort overrides, and session storage in Codex's native format for manual resumption with `codex resume`.
Cognium trust score
95%
Tier
Verified
Composite of vulnerability cleanliness, spec conformance, provenance, stability, and usage signals — scanned and weighted by Cognium. Human and agent signals are tracked separately.
Last scanned 2026-09-19.
Returns 7 tools: search_skills, get_skill, list_leaderboard, get_trust_breakdown, resolve_composition, plus the ChatGPT-connector search and fetch. Every tool is annotated read-only.
Resolve this skill directly via MCP tools/call get_skill.