agentleFS
ai-agent-assemblyai-agent-assembly/agent-assembly/.claude/skills/documentation-experience/SKILL.md

documentation-experience

Skill4 starsChanged 3 months agoApache-2.0View on GitHub

The company-common capability for designing and maintaining product documentation/communication that is immediately understandable at the surface, progressively deep when needed, persona-aware, source-grounded, and genuinely usable. Company-common per governance/engineering/agent-skill-architecture.md (HORO-969) §1. Simple first. Deep on demand. Never shallow when the user needs an answer. For landing/entry surfaces specifically: Within roughly five seconds, the intended visitor should understand what this product/project is, why it exists, why they might care, and the next useful action. This is a heuristic/Design QA target, not a…

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Don't force Layer 0 orientation content to carry Layer 3 reference detail, or vice versa.

    line 77

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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
<!-- horonom:generated --><!-- Source: horonomy/.github agents/skills/documentation-experience/SKILL.md. Do not hand-edit — rerun `python3 agents/common/project_skills.py`. -->

SKILL.md — documentation-experience

Purpose

The company-common capability for designing and maintaining product documentation/communication that is immediately understandable at the surface, progressively deep when needed, persona-aware, source-grounded, and genuinely usable. Company-common per governance/engineering/agent-skill-architecture.md (HORO-969) §1.

Core principle

Simple first. Deep on demand. Never shallow when the user needs an answer.

For landing/entry surfaces specifically:

Within roughly five seconds, the intended visitor should understand what this product/project is, why it exists, why they might care, and the next useful action.

This is a heuristic/Design QA target, not a literal stopwatch metric.

Type

Auto-used, no manifest.yaml — deliberately, matching shell-development and the other no-manifest company-process skills. Per governance/engineering/agent-skill-architecture.md §3.4, applicability here is broader than any single stack signal: it applies to any repo with user-facing documentation, rendered UI or not (a CLI with a real README gets it same as a web product) — so resolve_applicable_skills() correctly treats it as always applicable, with no Cargo.toml-style gate to write.

When to use

Authoring or materially rewriting any documentation/communication surface: a marketing/product homepage, a technical docs homepage, a README, user/operator/integrator docs, or contributor/developer docs — and when classifying Docs Impact for a material implementation/product/design change.

When NOT to use

  • Inventing product capability truth. This skill presents a product's Product Truth/Constitution — it never invents or overrides it (source of truth stays product-owned, per HORO-969 §1's ownership boundary).
  • As a substitute for product-validation's golden user-journey QA or design-qa's rendered-surface verification — see Composition below.

Composition

  • engineering-loop (routing only) — its Explore → Narrow → Validate → Escalate → Full Gate loop applies when iterating on docs source, scripts, or link/snippet checks; not redefined here.
  • design-qa (agents/skills/design-qa/SKILL.md) — marketing/docs entry surfaces are checked via its five-second-comprehension dimension; a marketing homepage must pass before its docs change is complete, a technical docs homepage is instead checked for task/audience navigation. Not duplicated here.
  • product-validation (agents/skills/product-validation/SKILL.md) — docs-driven user journeys (documentation scenario -> product-validation scenario -> actual supported product path) are a validation track owned there; a docs claim an E2E can't complete is a quality failure to track, not "just docs."

Routing decision

  1. Identify the surface and its communication job — see references/surface-communication-jobs.md. Don't write generic docs copy; know which of the five profiles applies before drafting.
  2. Place content on the right layer — see references/progressive-information-architecture.md's 4-layer model. Don't force Layer 0 orientation content to carry Layer 3 reference detail, or vice versa.
  3. Ground every claim in current source of truth — code/CLI/API/UI, Product North Star/Constitution, current release/deployment status, Jira/ADR where it owns the decision. Label maturity using the fixed five-term vocabulary in governance/engineering/docs-scenario-quality.md (SHIPPED/PARTIAL/PLANNED/INTERNAL-ONLY/NOT IMPLEMENTED).
  4. Use the semantic admonition system for secondary detail — see references/admonition-system.md. Map onto the docs site's own native callout primitive; never invent new rendering.
  5. Classify Docs Impact for the change per references/docs-impact-gate.md before considering delivery complete.
  6. Verify per governance/engineering/docs-scenario-quality.md's pre-ship checklist and independent-review expectation, composing design-qa/product-validation where the surface warrants it.

Generalizes, does not fork

Generalizes governance/engineering/docs-scenario-quality.md (the HORO-638 distillation: scenario-first expectation, maturity vocabulary, ground-truth reconciliation, pre-ship checklist, independent review) and HORO-820's Docs Impact gate into one company-wide capability — it does not restate HORO-638's canonical template or fork a second, drifting model. Read the governance doc for the process invariants; this skill supplies the operational how.

References and example

  • references/progressive-information-architecture.md — Layer 0–3 model with a worked example.
  • references/surface-communication-jobs.md — the five surface profiles and persona/role-driven guidance.
  • references/admonition-system.md — the eight semantic admonition types.
  • references/docs-impact-gate.md — the generalized Docs Impact gate.
  • examples/homepage-five-second-rewrite.md — a homepage failing five-second comprehension, diagnosed and rewritten to pass.

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.