A thin, auditable MCP server wrapping the Codex CLI (codex exec). Same tools as the deprecated `codex mcp-server`; forward-compatible with future CLI flags via extra_args passthrough.
Wellknown found it in public sources; nobody has proven control of it yet. Claiming takes one click if the repository is under your GitHub account, or a small file on your domain otherwise. Verified owners get the badge, 15-minute checks, status alerts, edits that outrank crawled data, and a ranking boost.
Agents can do it too: POST https://wellknown.network/api/v1/claims with {"agent":"codex-cli-mcp-slim","method":"well_known_file"} — machine-readable steps at claim.json, guide at /docs/claim.
Everything here was measured by our prober or read from a registry. Nothing is self-reported.
Attributed to the source that supplied each field. Treated as claims, not facts.
# codex-cli-mcp-slim A thin, auditable [MCP](https://modelcontextprotocol.io) server wrapping the [Codex CLI](https://github.com/openai/codex) (`codex exec`). [](https://pypi.org/project/codex-cli-mcp-slim/) [](https://pypi.org/project/codex-cli-mcp-slim/) [](https://opensource.org/licenses/MIT) [](https://github.com/tksfjt1024/codex-cli-mcp-slim/actions/workflows/ci.yml) ## Why `codex mcp-server`, the command that let other MCP clients call Codex, is deprecated, and its removal has been merged upstream ([openai/codex#42993](https://github.com/openai/codex/pull/42993)): releases up to 0.153.x still ship it, later ones will not. Its replacement, the Codex app server, speaks its own JSON-RPC protocol rather than MCP. This server keeps the old integration point alive: it exposes the same two tools, `codex` and `codex-reply`, and runs `codex exec` underneath. `codex exec` is the Codex CLI's non-interactive mode: one prompt in, the agent works on its own, one final message out. When you add an MCP server to your AI coding tool, every prompt and code snippet you send flows through that wrapper. Most CLI-wrapping MCP servers are small, individually maintained packages, and recent supply-chain incidents (`xz-utils`, `postmark-mcp`, the npm `chalk`/`debug` compromise) show that "small and useful" is not the same as "safe to trust blindly." This project takes the opposite stance: instead of asking you to trust it, it tries to be **easy to audit**. - **Single file** — the whole server is `src/codex_cli_mcp_slim/server.py`, readable end-to-end in one sitting - **One third-party dependency** (`mcp`) — minimal supply-chain surface - **Faithful CLI mapping** — every ty…
Mapped onto the structured taxonomy from declared text and observed tool names. Confidence shown for derived entries.
Every source is kept verbatim. Field changes are logged as events.