agentleFS
kala-projectkala-project/acp-kit/docs/agents.md

acp-kit

AGENTS.md2 starsChanged 7 months agoMITView on GitHub

ACP Kit ships six built-in AgentProfile constants. Each one is a small object describing how to spawn the corresponding ACP-capable CLI. You can use them directly, spread them to override one field, or write your own AgentProfile literal for any other agent that speaks ACP over stdio. Built-in profiles spawn the CLI binary from PATH instead of npx <package>@latest. This avoids an npm registry/version-resolution step on every agent launch; install the corresponding npm package globally or override command / args…

  • Reads credentials

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 1

    1. AcpKit/acp-kit★ 2100% same

    Edited · 0

    None found.

    Wrote their own version · 0

    None found.

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.

    Of the AGENTS.md files we read, the 31,774 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.

  • In this repositoryOne other file it gives its agents.

The file

We split this agents.md into the rules it gives an agent

1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use filterStdoutLine for agents that print non-protocol lines on the same channel as JSON-RPC frames.

    line 182

≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
agents.md

Supported ACP Agents

ACP Kit ships six built-in AgentProfile constants. Each one is a small object describing how to spawn the corresponding ACP-capable CLI. You can use them directly, spread them to override one field, or write your own AgentProfile literal for any other agent that speaks ACP over stdio.

Built-in profiles spawn the CLI binary from PATH instead of npx <package>@latest. This avoids an npm registry/version-resolution step on every agent launch; install the corresponding npm package globally or override command / args if you prefer a pinned local wrapper. When the primary command is not found on PATH, built-in profiles include fallbackCommands that attempt npx --yes <package>@latest automatically so first-run package installation cannot hang on npm's confirmation prompt.

import { createAcpRuntime, ClaudeCode } from '@acp-kit/core';

await using acp = createAcpRuntime({ agent: ClaudeCode, host });

The runtime forwards every standard ACP feature (session/load, setMode, setModel, session/list, file system, terminal capabilities) to whichever agent advertises it during initialize. Read acp.agentCapabilities after the first session is created to see exactly what a given agent CLI version supports today.


Claude Code

import { ClaudeCode } from '@acp-kit/core';
FieldValue
idclaude-code
displayNameClaude Code
commandclaude-code-acp
args[]
startupTimeoutMs30000
LoginAnthropic API key in ANTHROPIC_API_KEY, or interactive Claude.ai login on first run
Upstream@zed-industries/claude-code-acp

Anthropic Claude Code via the Zed-maintained ACP wrapper.


GitHub Copilot

import { GitHubCopilot } from '@acp-kit/core';
FieldValue
idgithub-copilot
displayNameGitHub Copilot
commandcopilot-language-server
args['--acp']
startupTimeoutMs30000
LoginGitHub OAuth device flow on first run; token cached by the language server
Upstream@github/copilot-language-server

GitHub Copilot's language server in ACP mode. Requires an active Copilot subscription on the signed-in GitHub account.


Codex CLI

import { CodexCli } from '@acp-kit/core';
FieldValue
idcodex-cli
displayNameCodex CLI
commandcodex-acp
args[]
startupTimeoutMs30000
LoginOpenAI API key in OPENAI_API_KEY
Upstream@zed-industries/codex-acp

OpenAI Codex CLI via the Zed-maintained ACP wrapper.


Gemini CLI

import { GeminiCli } from '@acp-kit/core';
FieldValue
idgemini-cli
displayNameGemini CLI
commandgemini
args['--experimental-acp']
startupTimeoutMs30000
LoginGoogle API key in GEMINI_API_KEY, or gcloud auth for ADC
Upstream@google/gemini-cli

Google Gemini CLI in experimental ACP mode. The --experimental-acp flag means upstream may break compatibility between minor versions; pin a specific CLI version in production.


Qwen Code

import { QwenCode } from '@acp-kit/core';
FieldValue
idqwen-code
displayNameQwen Code
commandqwen
args['--acp', '--experimental-skills']
startupTimeoutMs30000
LoginAlibaba DashScope API key in DASHSCOPE_API_KEY (or the CLI's interactive setup)
Upstream@qwen-code/qwen-code

Alibaba Qwen Code in ACP mode with the experimental skills feature enabled.


OpenCode

import { OpenCode } from '@acp-kit/core';
FieldValue
idopencode
displayNameOpenCode
commandopencode
args['acp']
startupTimeoutMs30000
LoginProvider keys configured via opencode auth login (multi-provider; reads from OpenCode's config)
Upstreamopencode-ai

OpenCode CLI in ACP mode.


Overriding a built-in profile

Spread the constant and replace any field. The runtime clones args and env internally, so the original constant stays untouched.

import { createAcpRuntime, ClaudeCode } from '@acp-kit/core';

await using acp = createAcpRuntime({
  agent: {
    ...ClaudeCode,
    env: { ANTHROPIC_API_KEY: process.env.ANTHROPIC_API_KEY! },
    startupTimeoutMs: 60000,
  },
  host,
});

startupTimeoutMs defaults to 30000 on built-in profiles but is not capped; raise it for slow first-run downloads or authentication-heavy startup flows.

Custom AgentProfile

Any binary that implements ACP over stdio works. The full interface:

interface AgentProfile {
  id: string;                       // stable id, surfaced in errors
  displayName: string;              // human-readable name
  command: string;                  // executable, resolved through PATH
  args: string[];
  fallbackCommands?: Array<{ command: string; args: string[] }>; // slower fallback launches, e.g. npx --yes
  env?: Record<string, string>;     // merged onto process.env
  startupTimeoutMs?: number;        // default 30000
  filterStdoutLine?: (line: string) => string | null; // drop chatty banners
}

const myAgent: AgentProfile = {
  id: 'my-agent',
  displayName: 'My Agent',
  command: 'my-agent-cli',
  args: ['--acp'],
};

await using acp = createAcpRuntime({ agent: myAgent, host });

Use filterStdoutLine for agents that print non-protocol lines on the same channel as JSON-RPC frames. Return null to drop a line, or a transformed string to keep it.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.