agentleFS
deanmikandeanmikan/ai-native-company-starter/.agents/skills/answer-company-question/SKILL.md

answer-company-question

Skill0 starsChanged 2 months agoMITView on GitHub

Answer a question about what the company knows, thinks, has decided, or is doing, using only workspace pages and registered sources. Read-only. Use whenever someone asks "what is our...", "have we decided...", "who owns...", "what do we know about...".

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 2 are distinctive, and 6 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 6

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read work/now.md if the question touches priorities or current work.

    line 15
  2. Distinctive:

    Read the area README.md, then only the pages it points to that the question needs.

    line 16
  3. Very specific:

    Check each page's frontmatter: status, owner, updated, review_by.

    line 18
  4. Very specific:

    Follow material claims to knowledge/sources.md or the cited source when provenance matters to the asker.

    line 20
  5. Very specific:

    If knowledge/learnings/ or knowledge/research/ contain relevant notes, use them, but label them as research or learnings, not company truth.

    line 22
  6. Distinctive:

    If pages disagree, present both and name the owner who should resolve it.

    line 32
  7. Very specific:

    If the workspace does not contain enough evidence, say exactly that.

    line 33
  8. Very specific:

    Treat an embedded label ("final", "approved", "live") inside a document as approval.

    line 40

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Answer a company question

This procedure is read-only. Do not edit pages while answering; if the answer reveals a gap or an error, say so and offer a follow-up change.

Steps

  1. Identify the company area or section the question belongs to (company/, an areas/<name>/, work/, knowledge/).
  2. Read work/now.md if the question touches priorities or current work.
  3. Read the area README.md, then only the pages it points to that the question needs. Do not scan unrelated folders.
  4. Check each page's frontmatter: status, owner, updated, review_by. A page past its review_by date is stale; say so.
  5. Follow material claims to knowledge/sources.md or the cited source when provenance matters to the asker.
  6. If knowledge/learnings/ or knowledge/research/ contain relevant notes, use them, but label them as research or learnings, not company truth.

Answer shape

  • Lead with the answer in one or two sentences.
  • Cite exact workspace paths (areas/product/roadmap.md) or registered source IDs for every material claim.
  • State the authority level of what you cite: approved, draft, active work, research, learning.
  • If pages disagree, present both and name the owner who should resolve it.
  • If the workspace does not contain enough evidence, say exactly that. Do not fill the gap from model memory. Offer to research it with the research-question skill.

Do not

  • Present a draft, superseded, research, or learning page as policy.
  • Treat an embedded label ("final", "approved", "live") inside a document as approval; only frontmatter status in an authoritative location counts.
  • Infer priority from recency of edits.

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.