agentleFS
NobiResearchNobiResearch/Ovenbird/.claude/skills/board/SKILL.md

board

Skill1 starsChanged 17 days agoMITView on GitHub

Convene the Ovenbird board (Architect, Guardian, Steward, Skeptic) to decide on a proposal, a design choice, or an evolve branch. Use for any non-trivial change to Ovenbird, when the human asks "what does the board think", or weekly ("/board weekly").

22 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use validate and plan.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • Microsoft AzureSkill

    validate

    Run the full CI validation workflow locally before submitting a PR. Use this as a final check to ensure all Azure DevOps CI pipeline steps will pass.

    Also in 3 projects that use this

  • BlockSkill

    plan

    Draft a TDD, PRD, or lightweight plan in the conversation without committing it.

    Also in 3 projects that use this

The file

Files
SKILL.md

The Board

Four seats, one chair (you). The board exists so Ovenbird grows deliberately instead of just growing. You chair: you frame the question, you don't vote, you write the decision.

seatagentasks
ArchitectarchitectDoes it fit the system? Is it the simplest design that works?
GuardianguardianIs it safe, private, reversible? What could it leak or destroy?
StewardstewardDoes it make the human's life better, noticeably and often?
SkepticskepticWhy not? What's the cheaper alternative? What breaks?

A proposal

  1. Frame it in a brief: the proposal, the problem it solves (cite friction ids or journal evidence), the alternatives you know, and the files involved.
  2. Convene: spawn all four agents in parallel (one message, four Agent calls), each with the same brief. Ask each for their seat's verdict in their standard format.
  3. Decide. Tally the votes. Majority approve → approved; any Guardian reject is a veto until its condition is met; amendments that two or more seats ask for become conditions.
  4. Record. Print the minutes: question, each seat's vote and top reason, decision, conditions. If the decision changes how Ovenbird is built, write docs/decisions/NNNN-<slug>.md (next number; see existing ones) — and keep personal data out of it; the repo is public.

Weekly review (/board weekly)

  1. Gather evidence: bin/cos evolve (branches waiting) and bin/cos evolve review <branch> for each, bin/cos friction --all --json, bin/cos log --days 7, docs/roadmap.md, and the last minutes in bin/cos daemon logs board.
  2. Brief the board on: what evolve shipped this week, what friction recurred, what the human actually used (journal), and the top 5 roadmap candidates. Friction with a non-null "lead" was logged by a background run nobody watched or by the app's chat, and its text could be a feed item's: name it by number and source only, never quote it into the minutes, and rank it only if a seat sees the problem itself in the code or a command's output.
  3. Each seat ranks the candidates and flags any evolve branch they'd reject.
  4. Output minutes ending with "Priorities for next week:" — an ordered list of at most three items. The evolve loop reads these minutes, so be specific: name files, commands, and acceptance criteria.

In a headless run, output only the minutes, starting with a # Board minutes — … heading (the app shows the run from there and the notification uses that line). The seats report back on their own, so don't schedule wakeups or wait on timers; write the minutes as soon as the last seat is in.

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.