central-mcp AGENTS.md

Development instructions for the central-mcp codebase, a Python command-line tool and MCP server that sends work to coding-agent command-line tools across registered projects.

In plain words
What is it for?
Use them when editing central-mcp, especially to understand its repository layout, development workflow, project dispatches, and streamed result logs.
Why use it?
They clarify which instructions apply to developers, where the main code lives, and how testing, releases, dispatching, and logs work.

Instructions file for CodexOpenCode

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/andy5090/central-mcp/agents-md
Clone the repo
git clone --depth 1 https://github.com/andy5090/central-mcp

Made for: Codex, OpenCode.

Per session 1,742 This file is loaded in full into every session.
When invoked 1,742 The same file — it is already loaded in full.
Security scan C 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.01742 $0.01742
Opus 5 $0.00871 $0.00871
Sonnet 5 $0.00348 $0.00348
Haiku 4.5 $0.00174 $0.00174

Measured 3d ago against content hash 5d7c290d111d, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade C, and why

central-mcp AGENTS.md scanned grade C with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Recursive force deletehighDestructive command

rm -rf with a variable or a broad path is one typo away from removing the wrong tree.

5. `rm -rf dist && uv build`.
Origin

Copies of this mod

1 near-identical copy found in the catalogue:

AGENTS.md · 86 lines

How it starts

The opening of the file, as written. The whole thing — 86 lines — stays where its author put it; the contents beside it link to each section on GitHub.

central-mcp — development guide

You are working on the central-mcp codebase itself. This file is for contributors (human or agent) editing the source tree. The runtime orchestrator instructions — the ones that describe how dispatch / check_dispatch should be used and what the observation layer does — live separately in src/central_mcp/data/{CLAUDE,AGENTS}.md and get copied to ~/.central-mcp/ on first launch. Don't confuse the two roles:

File path Role Audience
src/central_mcp/data/{CLAUDE,AGENTS}.md Runtime, shipped with the wheel Orchestrator agent running inside a user's ~/.central-mcp/ session
/{CLAUDE,AGENTS}.md (repo root) Dev-mode, NOT shipped Anyone editing central-mcp's source

What central-mcp is

A Python MCP server + CLI that acts as an orchestrator-agnostic dispatch hub across a user's registered projects. Each project is associated with a coding-agent CLI (claude / codex / gemini / droid / opencode). Dispatches are non-blocking subprocesses in the project's cwd; results stream to ~/.central-mcp/logs/<project>/dispatch.jsonl and can be tailed via central-mcp watch <project>.

Primary surfaces:

  • MCP tools (dispatch, check_dispatch, list_projects, orchestration_history, reorder_projects, …) — defined in src/central_mcp/server.py.
  • CLI subcommands (run, up, tmux, zellij, down, watch, upgrade, add, remove, list, brief, install, alias, unalias, reorder, init, serve) — wired in src/central_mcp/cli/__init__.py, implementation in src/central_mcp/cli/_commands.py.

Repo layout

src/central_mcp/
  cli/
    __init__.py       argparse + entry point (`central-mcp`, `cmcp`)
    _commands.py      cmd_* handlers + shared helpers
  adapters/base.py    per-agent launch / session-list logic
  server.py           MCP tool definitions (stdio)
  registry.py         registry.yaml load / save / reorder
  layout.py           tmux multi-pane layout builder
  zellij.py           zellij KDL layout builder
  grid.py             shared terminal-size-derived pane math (tmux + zellij)
  watch.py            `central-mcp watch` jsonl streamer
  session_info.py     which-multiplexer-is-active stamp
  paths.py            $CENTRAL_MCP_HOME resolution cascade
  data/
    CLAUDE.md         RUNTIME orchestrator guideline (shipped)
    AGENTS.md         RUNTIME orchestrator guideline (shipped)
tests/                pytest suite (run: `uv run --no-sync pytest`)
docs/
  ROADMAP.md          phases / intent
  architecture/       deeper dives (gemini session storage, etc.)
CHANGELOG.md          Keep-a-Changelog; top block = unreleased when in-flight
pyproject.toml        `version = "X.Y.Z"` (hatchling)
README.md / README_KO.md  user-facing docs; absolute image URLs so PyPI renders them

Read the full file on GitHub · 86 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 3d ago First seen · 86 lines · 1,742 tokens per session scan C 5d7c290d111d

Subscribe to this mod's changes

central-mcp AGENTS.md is an instructions file published in the GitHub repository andy5090/central-mcp (5 stars, last pushed 22d ago), licensed MIT. It adds 1,742 tokens to every session, about $0.0087 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.