Skip to content
Back to skills

Intuitive Doc

ASecurity

Maintain README.md, ARCHITECTURE.md, STATUS.md, and docs/human/** when they drift or need cleanup.

  • 48 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added October 11, 2026
ai-agentsgodocumentation

Security analysis

A100/100

Pro scans all 3 files and shows the line behind each finding

Scanned October 11, 2026

npx -y skills add MiaoDX/intuitive-flow --skill intuitive-doc --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Intuitive Doc?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Intuitive Doc
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/miaodx-intuitive-doc/badge)](https://www.skillsdirectory.com/skills/miaodx-intuitive-doc)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: intuitive-doc
description: Maintain README.md, ARCHITECTURE.md, STATUS.md, and docs/human/** when they drift or need cleanup.
disable-model-invocation: true
---

# Intuitive Doc

Use this skill to keep the human-facing documentation surface small, current,
and reviewable. The default human surface is:

- `README.md`;
- `ARCHITECTURE.md`;
- `STATUS.md`;
- `docs/human/**`.

Planning logs, generated evidence, retrospectives, ADR detail, execution
artifacts, scratch notes, and agent-only runbooks should stay outside the human
surface unless a human doc intentionally promotes them.

For plan and task-state placement, follow [plan selection](../_shared/references/plan-paths.md).
Preserve the repo's existing canonical plans and lifecycle layout. When adopting
Intuitive defaults, use a flat plan directory and keep current execution state
in the task-owned resume surface. A plan dashboard is an agent navigation aid.

This compact entrypoint preserves full mode-specific guidance in
[detailed guidance](references/detailed-guidance.md). Read it for detailed
audit/update/cleanup criteria, perspective levels, documentation standards, or
full output format.

## When To Activate

Use this skill when:

- root docs drift from current code, commands, or active plans;
- humans need a concise orientation after AI-heavy development;
- implementation changed public behavior, command surfaces, setup, architecture,
  or status;
- generated/planning/agent evidence has leaked into human docs;
- the user asks for documentation cleanup, human docs, README/architecture/status
  updates, or post-implementation doc alignment.

## Modes

| Mode | Use when | Output | Redirect when |
| --- | --- | --- | --- |
| Audit | The user asks whether docs are current, bloated, or misplaced. | Findings, recommended changes, parked items, verification. | The issue is agent guidance, tests, or code layout. |
| Update | Current behavior, commands, architecture, or status changed. | Focused doc edits tied to current source evidence. | The requested change is only planning/evidence archival. |
| Cleanup | Human docs contain stale, duplicated, or over-detailed material. | Moved, archived, shortened, or removed docs with links preserved. | The material belongs in agent guidance or execution plans. |
| Guard | A proposed doc change needs placement review. | Placement decision and reason. | The user already selected the doc target and asks to edit. |

Name the selected mode in one sentence before non-trivial audits or edits, and
mention a better owner only when one exists.

## Human Surface Rules

- `README.md` orients and points to the current map; it should not become a
  manual.
- `ARCHITECTURE.md` owns subsystem contracts, data flow, extension points, and
  proof boundaries.
- `STATUS.md` owns current state, supported commands, next maintenance focus,
  and known blockers.
- `docs/human/**` owns human-facing detail that would bloat root docs.
- Agent runbooks belong in `docs/agents/**` or skill references unless humans
  need them as project truth.

## Output

For audit/report-only work, return:

```text
Human docs status:
Findings:
Recommended changes:
Parked items:
Verification:
```

For edits, report changed files, the current-truth source used, verification
run, and any docs intentionally left unchanged.

## Stop Conditions

Stop when the small human surface lets a human answer what the project is, how
to run or verify it, what is current, what changed, and where deeper detail
lives. Once routing, truth, and links are correct, stop; wording polish is not the goal.

Files in this skill

  • SKILL.md3.5 KB
  • agents/openai.yaml43 B
  • references/detailed-guidance.md6.4 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…