Mola-maker/mstack / skills/m-team
m-team
VerifiedEvolving agent team for open-ended, large-budget goals; a supervisor redesigns roles each round. Use for 组队/一直优化/卡在局部最优.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 11.4 KB |
| references/dsh-team.md | markdown | 3.5 KB |
| references/role-catalog.md | markdown | 5.1 KB |
| references/supervisor-checklist.md | markdown | 4.6 KB |
| references/team-config-template.md | markdown | 2.9 KB |
references/team-config-template.md
View on GitHub# Team config template (m-team)
Copy this file to `.mstack/team/<slug>/v<N>.md` at the start of every round and fill in every field. One file per round; never edit a previous round's file — supersede it.
```markdown
# Team v<N> — <slug>
- **Round:** <N>
- **Objective (from m-goal):** <one testable sentence; goal file: .mstack/goals/<slug>.md>
- **Metric:** <name, unit, how measured (exact command)>
- **Metric at round start:** <value + evidence>
- **Round budget:** <tokens / $ / wall-clock — the cap the whole team shares>
- **Noise band (write before scoring):** <spread from earlier measurements on unchanged code, or the smallest delta the command resolves; mark "weak" if you have no prior spread. Do not fill this from this round's results>
- **Solo baseline:** <same cap as Round budget, one agent, not a role; or "this round is already solo">
- **Round threshold:** <the metric value that makes this round a success>
- **Diagnosis from last supervisor round:** <knowledge gap / exhausted search / discipline problem — one line why> (v1: "initial design")
## Gates (any failure → candidate scores 0)
1. <gate: exact command or check, expected result>
2. …
## Held-out checks (checked, never scored)
1. <check + command>
2. …
## Coordination shape
<solo | pipeline | parallel branches> — <one line: who hands what to whom, in which order or in parallel. The solo baseline is not a role in the list below; it runs beside a pipeline or parallel team and writes only its own workspace>
## Roles
### Role 1: <name>
- **Duty:** <what this role exists to do this round>
- **May change:** <files / directories / branch>
- **Must not touch:** <explicit forbidden areas>
- **Delivers (evidence):** <files, diffs, measurements, failing cases — concrete artifacts>
- **Writes to:** <file path or branch name>
- **Reads first:** `SYNTHESIS.md` (mandatory), plus <round-specific inputs>
### Role 2: …
(repeat; add or drop roles freely — see role-catalog.md)
## Knowledge placement
- **Shared log:** `.mstack/team/<slug>/LEARNINGS.md` (append-only, each entry with evidence)
- **Actionable conclusions:** `.mstack/team/<slug>/SYNTHESIS.md`
- **This round's artifacts:** <directory or branches where this round's outputs land>
- **Round-specific notes:** <which role writes which extra file, if any>
## Round close (filled in by the supervisor)
- **Metric after round:** <value + evidence>
- **Solo metric / team-best metric:** <values, or "n/a, this round was solo">
- **Gain vs noise band:** <improvement of team over solo in the metric's success direction; inside band → stop teaming and hand back to m-goal>
- **Label map (fill after scoring, never send to the scorer):** <label → role or solo>
- **Candidates scored 0:** <list + failed gate>
- **Held-out results:** <pass/fail per check>
- **Handed to supervisor checklist at:** <timestamp>; next version: v<N+1> or stop
```
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |