COMMS.md Creator

Create a COMMS.md — a structured, queryable document expressing someone's communication preferences for humans and agents. Use when: (1) someone wants to art...

Security Scan

This instruction-only skill cleanly guides a conversational interview and generates a COMMS.md document; its requirements and instructions are consistent with that purpose and it doesn't request credentials or install code.

Purpose & Capability

Name, description, and provided files (template + example + SKILL.md) align with the stated goal of generating a COMMS.md. No unrelated binaries, env vars, or configs are requested.

Instruction Scope

Runtime instructions are a step-by-step interviewing and drafting workflow limited to asking questions, filling a provided template, reviewing with the user, and asking where to save the document. There are no instructions to read arbitrary system files, access credentials, or transmit data to hidden endpoints.

Install Mechanism

No install spec or code files are provided; this is instruction-only, so nothing is written to disk or downloaded by the skill itself.

Credentials

The skill requests no environment variables, credentials, or config paths. The guidance to ask where to save the document mentions external targets (personal website, Notion, Obsidian) but does not itself require or access those services' credentials.

Persistence & Privilege

always is false and the skill has normal user-invocable/autonomous-invocation defaults. The skill does not request persistent agent-level privileges or attempt to modify other skills or system settings.

Assessment

This skill appears coherent and low-risk: it only contains instructions and templates for an interview that produces a COMMS.md. Before installing/using, consider privacy and placement: the generated document may include sensitive preferences or personal examples, so be deliberate about where you (or an agent) save or publish it (personal website vs private vault). The skill itself does not request credentials, but if you instruct your agent to save the COMMS.md into a third-party service (Notion, GitHub, etc.), that will involve separate credentials and permissions — verify those flows and consent before allowing the agent to store or publish the document. If you plan to pair this with automations (e.g., comms-md-reader), confirm access controls so the COMMS.md is shared only with intended parties.

Generate a structured communication preferences document through guided conversation.

What COMMS.md Is

A queryable personal document that expresses how someone communicates: their natural style, channel preferences, availability rhythms, async voice, and interaction protocols. Designed to be read by both humans and agents before initiating contact.

Workflow

1. Orient

Explain the concept briefly:

"A COMMS.md is a structured doc that captures how you communicate — your style, when you're available, which channels you prefer, how you write in different contexts. Think of it as a user manual for reaching you. Other people (or their agents) can reference it before getting in touch."

2. Interview

Walk through these areas conversationally. Don't dump all questions at once — do 2-3 per turn, adapt based on answers, skip what's clearly not relevant.

Style & strengths (start here — it's introspective and sets the tone):

• What comes naturally when you communicate? (e.g. brevity, storytelling, humor, directness)
• What requires effort? (e.g. small talk, follow-through, emotional labor, context-switching)
• Where does communication break down for you?

Collaboration model:

• What kind of people do you work best with?
• How do you prefer to set up working relationships?

Weekly rhythm:

• Walk through their week: energy, availability windows, protected time
• Which days are meeting-heavy? Which are deep work?

Sync philosophy:

• What are calls for? What are they NOT for?
• How tactical vs. strategic do calls get?

Channel preferences:

• For each situation type (urgent, professional, casual, etc.): what channel, what timing?
• Role of email vs. text vs. voice notes vs. calls?
• How does closeness change channel choice?
• Notification habits, response triage, focus mode patterns

Async voice (this section often needs the most drawing out):

• How do they write to close friends vs. professional contacts?
• Capitalization, punctuation, emoji habits?
• What do they never do in writing? (anti-patterns)
• How do they handle outreach to new people? Re-engaging after silence?
• Do they have a gap between how competent they are and how warm they read? How do they bridge it?

Boundaries:

• What's the fastest way to annoy them?
• What should people never do?

3. Draft

Read references/comms-template.md for the output structure. Generate the COMMS.md using the template, filling in sections from the interview. Use the person's own words and phrasing where possible.

For reference on what a completed COMMS.md looks like, see references/example.md.

4. Review and Iterate

Present the draft and ask what feels off, what's missing, what's too specific or too vague. Expect 1-2 revision rounds. Common adjustments:

• Tone too formal or too casual for how they actually talk
• Missing a channel or context they care about
• Weekly rhythm needs more nuance
• Anti-patterns section needs real examples

5. Place the Document

Ask where they want it saved. Common locations:

• Personal website (for public/professional use)
• Notes app or vault (Obsidian, Notion, etc.)
• Workspace docs (for team use)

Related Skills

• comms-md-reader — the companion skill for reading and adapting to someone else's COMMS.md when drafting outreach

Guidelines

• Use their voice. The doc should sound like them, not like a template. Mirror their register.
• Earn specificity. Generic preferences ("I prefer email for professional stuff") are less useful than specific ones ("Email for intros beyond close friends and anything they'll need to find later").
• Skip irrelevant sections. Not everyone has a weekly rhythm to document or sync philosophy to articulate. Leave out what doesn't apply rather than filling with filler.
• The Async Voice section is the highest-value section for agent consumption. Spend extra time here. This is what another agent reads to calibrate tone when drafting a message to this person.
• Version the output. Include Version 0.1 at the bottom so they can track iterations.