Guide · Agent skill
How to use the Board Meeting Protocol agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The board-meeting skill is a set of instructions that tells an AI agent to run a simulated board meeting on a strategic question. It makes the agent play several executive roles, such as CEO, CFO, CTO and CMO, who each give a view without seeing the others first, then produces a synthesis for you to approve or reject. This guide explains what the skill does, how to install it, how to trigger it and what it reads and writes on your machine.
Board Meeting Protocolalirezarezvani/claude-skills · MIT Download the skill Quick start
- 1Download and unzip. Download the skill ZIP from this page and unzip it to get the board-meeting folder.
- 2Place the folder. Move it to ~/.claude/skills/ for personal use or to .claude/skills/ in a project. Alternatively, upload it under Settings → Capabilities → Skills in the Claude apps.
- 3Install the C-level agent set. Make sure the claude-skills C-level agent set is installed, since the skill depends on its command and on agent-protocol/SKILL.md.
- 4Prepare company context. Write relevant facts to ~/.claude/company-context.md so the roles reason from your situation.
- 5Invoke a meeting. Run /c-level-agents:boardroom with your strategic question, or describe the meeting in natural language.
- 6Review the synthesis. Approve, modify, reject or ask a follow-up when the agent pauses for founder review.
- 7Check the saved records. After approval, look in ~/.claude/decisions/ for the transcript, the approved record and the index.
What the board-meeting skill makes the agent do
An agent skill is a folder containing a SKILL.md file of instructions that an agent such as Claude Code loads when a request matches it. This one implements a structured 6-phase deliberation: context loading, independent C-suite contributions, critic analysis, synthesis, founder review and decision extraction.
The isolation rule is central. Each role contributes before it sees other roles' outputs, which the skill describes as preventing groupthink. A critic step then looks for suspicious agreement, shared unvalidated assumptions and missing voices, and the agent presents a synthesis that you approve, modify, reject or question.
- Phase 1: load your company context and approved past decisions, then confirm the agenda and roles
- Phase 2: each activated role gives up to 5 key points, each tagged VERIFIED or ASSUMED with a confidence level
- Phase 3: a critic reviews all contributions at once
- Phase 4: synthesis in a fixed output format
- Phase 5: the agent stops and waits for your review
- Phase 6: approved decisions and transcripts are saved to local files
Install the board-meeting skill
Download the ZIP of the skill from this page and unzip it. For a personal install, place the folder in ~/.claude/skills/. For a single project, place it in that project's .claude/skills/ folder.
Alternatively, in the Claude apps, go to Settings → Capabilities → Skills and upload the skill there.
The skill refers to a /c-level-agents:boardroom command and to a file called ../agent-protocol/SKILL.md. That file is not included in the reviewed files, so install the claude-skills C-level agent set alongside this skill.
- Personal install: ~/.claude/skills/board-meeting/
- Project install: .claude/skills/board-meeting/ inside your project
- App install: Settings → Capabilities → Skills
Trigger the skill with a request
The skill activates when you invoke /c-level-agents:boardroom followed by a topic, when you say you are calling a board meeting, or when you ask for structured multi-perspective executive deliberation on a strategic question. The skill's example is a market question.
If a meeting was auto-closed as pending review, the skill says you can reopen it with /c-level-agents:boardroom resume.
- Use the command form: /c-level-agents:boardroom Should we expand to Spain in Q3?
- Use a natural request: Call a board meeting on whether we should raise our prices this year.
- Ask for deliberation: I want structured executive input on whether to hire a VP of Engineering before our next funding round.
/c-level-agents:boardroom Should we expand to Spain in Q3?Choose the right roles and prepare your inputs
The Chief of Staff role selects only the roles relevant to the topic rather than all fourteen each time. For example, a market expansion question activates CEO, CMO, CFO, CRO and COO, while a pricing question activates CMO, CFO, CRO and CPO.
The skill reads ~/.claude/company-context.md at the start. Writing a short, accurate version of that file helps the roles reason from your real facts. You can also paste or save an agenda using the agenda template in the skill's templates folder.
- Write your company facts into ~/.claude/company-context.md before the meeting
- Name the decision in one sentence, since the synthesis begins with a 'Decision Required' line
- Confirm the proposed agenda and roles when the agent asks, as the skill waits for your confirmation
Review, approve and find your saved records
At the review stage the agent shows the synthesis with four options: approve, modify, reject or ask a follow-up. The skill states that your corrections override the agent's proposals. After approval, the agent writes a full transcript to ~/.claude/decisions/raw/ and an approved record to ~/.claude/decisions/approved/, and adds it to an index file, decisions.md. Rejected proposals are marked [DO_NOT_RESURFACE].
Future meetings load only the approved records, never the raw transcripts, which the skill says prevents hallucinated consensus.
- Raw transcripts: ~/.claude/decisions/raw/YYYY-MM-DD-<slug>.md
- Approved records: ~/.claude/decisions/approved/YYYY-MM-DD-<slug>.md
- Index: ~/.claude/decisions/approved/decisions.md
What people use it for
- Stress-testing a market expansion decision before committing budget
- Getting a structured view on pricing from marketing, finance, sales and product perspectives
- Deciding whether to hire a specific role, with CEO, CHRO, CFO and COO input
- Keeping a running log of past strategic decisions that later meetings can reference
Frequently asked questions
How do I install the board-meeting skill?
Download the ZIP from this page and unzip it into ~/.claude/skills/ for personal use, or into a project's .claude/skills/ folder. In the Claude apps you can instead upload it under Settings → Capabilities → Skills.
What do I need to run it?
Claude Code or a compatible agent with the claude-skills C-level agent set installed. The skill refers to a /c-level-agents:boardroom command and to ../agent-protocol/SKILL.md, which is not included in the reviewed files. It also needs write access to ~/.claude/company-context.md and ~/.claude/decisions/.
Does the skill run scripts or use the network?
According to the review, it runs no scripts and uses no network. It does write files to your home folder, specifically ~/.claude/decisions/, where approved decisions and transcripts are saved.
Is the board-meeting skill free?
The skill is published under the MIT licence. The source does not mention pricing, so check the directory page and the repository for any other terms.
Can I use this instead of a real lawyer, accountant or doctor?
No. The output is simulated role-play by an AI, not professional advice. For high-stakes legal, financial or medical decisions, the reviewed assessment advises against relying on it.
Why does my board meeting not start or fail to find the command?
The skill depends on the claude-skills C-level agent set, including the /c-level-agents:boardroom command and ../agent-protocol/SKILL.md. If those are missing, check that the full set is installed alongside this skill, and that the skill folder sits in one of the documented locations.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use this skill if you are a founder or small-business owner who wants a structured stress test of a strategic decision, or you already use the claude-skills C-level agent set and want a decision log. Pick something else if you need real stakeholder input, since the customer and front-line voices are only what the model imagines, or if you do not want the agent writing files into ~/.claude/decisions/.
Download the skill