mate

Plugin mate from marketplace agentmate in naldomadeira/agentmate. Commands read from the repo's manifest — 12 days ago.

Skills (18)

• codex

  Route a task or question to OpenAI Codex (quick questions, reviews, research, plans, scoped implementation or a team lead) by picking the matching AgentMate tool. Use when the user says /mate:codex, 'ask codex', 'check with codex', 'hand this to codex', 'pergunte ao codex'.

• agy

  Route a task or question to the Google Antigravity CLI (agy), experimental (quick questions, reviews, research, plans, scoped implementation or a team lead) by picking the matching AgentMate tool. Use when the user says /mate:agy, 'ask agy', 'check with agy', 'hand this to agy', 'pergunte ao agy', o

• ask

  Ask Codex, Claude, Gemini or Antigravity (the last two experimental) a direct question and get the answer in the same turn. Use when the user says /mate:ask, 'ask codex', 'ask claude', 'ask gemini', 'ask agy', 'pergunte ao codex/claude', or wants a quick second opinion from the other model.

• claude

  Route a task or question to Claude Code (quick questions, reviews, research, plans, scoped implementation or a team lead) by picking the matching AgentMate tool. Use when the user says /mate:claude, 'ask claude', 'check with claude', 'hand this to claude', 'pergunte ao claude'.

• copilot

  Route a task or question to the GitHub Copilot CLI, experimental (quick questions, reviews, research, plans, scoped implementation or a team lead) by picking the matching AgentMate tool. Use when the user says /mate:copilot, 'ask copilot', 'check with copilot', 'hand this to copilot', 'pergunte ao c

• crossreview

  Have one agent implement a scoped change and the other agent review it, looping on the findings without you relaying anything by hand. Use when the user says /mate:crossreview, 'cross-review', 'cross review this', 'revisão cruzada', 'implementa e o outro revisa', 'codex implementa e o claude revisa'

• cursor

  Route a task or question to the Cursor CLI, experimental (quick questions, reviews, research, plans, scoped implementation or a team lead) by picking the matching AgentMate tool. Use when the user says /mate:cursor, 'ask cursor', 'check with cursor', 'hand this to cursor', 'pergunte ao cursor', or n

• delegate

  Delegate a task to another agent CLI (codex, claude, gemini or agy; the last two experimental) as a background job, covering second opinions, reviews, research and scoped implementation. Use when the user says /mate:delegate, 'hand this to codex', 'delegue ao claude', or wants an independent opinion

• gemini

  Route a task or question to Google Gemini CLI, experimental (quick questions, reviews, research, plans, scoped implementation or a team lead) by picking the matching AgentMate tool. Use when the user says /mate:gemini, 'ask gemini', 'check with gemini', 'hand this to gemini', 'pergunte ao gemini'.

• grok

  Route a task or question to xAI Grok Build, experimental (quick questions, reviews, research, plans, scoped implementation or a team lead) by picking the matching AgentMate tool. Use when the user says /mate:grok, 'ask grok', 'check with grok', 'hand this to grok', 'pergunte ao grok', or names xAI G

• implement

  Delegate a scoped code change to Codex, Claude, Gemini or Antigravity (the last two experimental) with write access, with acceptance criteria and verification. Use only when the user says /mate:implement, 'have codex implement this', 'peça ao claude para implementar', or otherwise explicitly authori

• init

  Set up AgentMate in a repository by adding or refreshing a short marked block in AGENTS.md and CLAUDE.md that tells every agent the /mate: commands and the rules for jobs it receives. Use when the user says /mate:init, 'set up agentmate in this repo', 'configurar o agentmate', 'prepare this repo for

• jobs

  List, observe, collect or cancel background jobs started through AgentMate. Use when the user says /mate:jobs, 'is the job done', 'list my jobs', 'cancel that job', 'ver os jobs', or needs a job id or a stored result.

• plan

  Get an implementation plan from Codex, Claude, Gemini or Antigravity (the last two experimental), or a critique of a plan you already have, with files, risks and verification steps. Use when the user says /mate:plan, 'plan this with codex', 'challenge this plan', 'planeje com o claude', or before a

• research

  Investigate a technical topic with Codex, Claude, Gemini or Antigravity (the last two experimental) and get findings, compared options and a recommendation with evidence. Use when the user says /mate:research, 'research this with codex', 'pesquise com o claude', or needs a library choice, root-cause

• review

  Get an independent code review of a diff, branch, files or a plan from Codex, Claude, Gemini or Antigravity (the last two experimental), with findings ranked by severity. Use when the user says /mate:review, 'have codex review this', 'peça ao claude para revisar', or wants a second reviewer before m

• split

  Split a broad goal into independent parts that both agents work on in parallel, then cross-review each part and report how to integrate them. Use when the user says /mate:split, 'split the task', 'task splitting', 'split this between codex and claude', 'dividir a tarefa', 'divida o trabalho entre os

• teamlead

  Hand a broad objective to Codex, Claude, Gemini or Antigravity (the last two experimental) acting as team lead, which breaks it down and delegates pieces to the other model as child jobs. Use when the user says /mate:teamlead, 'run a team lead', 'let codex lead this', 'coloque o claude como líder',

Subagents (4)

• codex-researcher

  Research assistant backed by OpenAI Codex. Use to investigate a library, API, design trade-off or unfamiliar part of a codebase, or to hunt a root cause, when you want options compared and a recommendation with evidence rather than raw Codex output.

• codex-reviewer

  Independent code reviewer backed by OpenAI Codex. Use after a meaningful change, before a merge, or when the user wants a second reviewer on a diff, branch, set of files or plan. Returns verified findings ranked by severity, not raw Codex output.

• codex-teamlead

  Runs OpenAI Codex as a team lead for a broad, multi-part objective. Codex plans it, delegates pieces to Claude as child jobs, reviews them and reports back. Use when the work spans several tasks you want one coordinator to own, not for a single question or review.

• codex-teammate

  Codex-powered teammate for second opinions and general delegation. Use when you want OpenAI Codex to answer a question, sanity-check an approach, explain code or take a bounded task, and you want a synthesized answer back rather than raw output. For reviews use codex-reviewer, for investigations cod

MCP servers (1)

• agentmate

  runs: npx -y agentmate serve jobs

Hooks (2)

• UserPromptSubmit

  node "${CLAUDE_PLUGIN_ROOT}/hooks/user-prompt-submit.mjs"

• SessionStart

  node "${CLAUDE_PLUGIN_ROOT}/hooks/session-start.mjs"

Similar plugins