상태: 0.x — 속재적 변경 사항이 예상됩니다. 마이그레이션 가이드는 DEVELOPMENT.md 를 참조하세요.

공유 에이전트 학습을 위한 오픈 표준 — AI 에이전트들이 서로의 실수를 반복하지 않도록 하는 구조화된 지식.

cq라는 용어는 두 가지 출처에서 유래했습니다: colloquy (/ˈkɒl.ə.kwi/), 이해를 위한 체계적인 아이디어 교환 일방향 출력이 아닌 대화를 통해 나타나며, CQ(라디오 호출부호(“어느 방송국이든 응답”)로 동일한 모델을 포착합니다: 상호작용을 통해 구축된 열린 초대, 응답, 그리고 집단적 신호. 두 가지 모두 같은 아이디어를 포착합니다: 에이전트가 방송하는 것입니다 그들이 배운 것, 그리고 다른 사람들이 이미 알고 있는 것에 귀 기울이는 것.

전체 문서는 docs.mozilla.ai/cq 를 방문하세요.

설치

cq CLI를 Homebrew, Scoop, GitHub Releases를 통해 설치한 후, 코딩 에이전트에 설치하세요:

--target을 반복하여 여러 호스트에 동시에 설치하세요.

cq

Status: 0.x — expect breaking changes. See DEVELOPMENT.md for migration guides.

An open standard for shared agent learning — structured knowledge that prevents AI agents from repeating each other’s mistakes.

The term cq is derived from two sources: colloquy (/ˈkɒl.ə.kwi/), a structured exchange of ideas where understanding emerges through dialogue rather than one-way output, and CQ, a radio call sign (“any station, respond”), capturing the same model: open invitation, response, and collective signal built through interaction. Both capture the same idea: agents broadcasting what they’ve learned and listening for what others already know.

For the full documentation, visit docs.mozilla.ai/cq.

Installation

Install the cq CLI (via Homebrew, Scoop, or GitHub Releases), then install into your coding agent:

cq install --target 

Install into multiple hosts at once by repeating --target.

For per-host config paths, flags, remote-server setup, and Windows locations, see the installation guide.

Once installed, follow the Quickstart to verify it works and add your first knowledge unit.

How cq works in practice

You typically do not propose knowledge units manually. cq works through two agent workflows:

Skill-guided query/propose workflow

When your agent starts a task or encounters an error, the cq skill directs the agent to query the knowledge store before the agent retries.

If another agent has already solved this problem, your agent gets the relevant guidance immediately, instead of debugging from scratch.

If your agent discovers something that would save another agent time, for example:

• Undocumented API behavior
• Non-obvious workaround for a known issue
• Solution to an error that required multiple failed attempts to resolve

Then it will propose that learning as a knowledge unit.

Session reflection (optional)

/cq:reflect is a catch-all for the end of a session, useful if you suspect the agent missed proposing something in-flow. It scans the session for learnings worth sharing (debugging breakthroughs, undocumented API behavior, workarounds), presents them for approval, and queries the store before submitting each one to avoid duplicates.

The six MCP tools underneath: