Create a baby-step documentation plan (overview + domain deep dives for RAG). Asks doc language before writing. Use when planning repo docs or invoking /document-plan.
Pro scans all 2 files and shows the line behind each finding
Scanned 9/25/2026
npx -y skills add tibursocampos/agent-dev-toolkit --skill document-plan --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Document Plan?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/tibursocampos-document-plan-agent-dev-toolkit)More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.
---
name: document-plan
description: Create a baby-step documentation plan (overview + domain deep dives for RAG). Asks doc language before writing. Use when planning repo docs or invoking /document-plan.
---
## STOP - Read before ANY tool call
1. Read `{{GUARDRAILS_PATH}}`
2. Read `_shared/sdd-artifacts/SESSION.md`; load session-state for `$Cwd`
3. If the relevant gate is not approved: **STOP** - ask user **(pt-BR)** - do **NOT** Write/Shell
4. SDD/develop skills: after **ONE** step/task, **STOP** session - handoff only
5. This skill body is **English**; user-facing prompts may be **(pt-BR)**
### Step -1 - Gate check (report in chat before continuing)
```
Gate check:
[ ] guardrails.mdc read
[ ] SESSION.md read; session-state loaded
[ ] PIPELINE.md read (SDD skills only)
[ ] User confirmed current action (sim)
-> If any unchecked: STOP
```
---
# Skill: document-plan
## Trigger
Invoke when the user asks for: `/document-plan`, `plan repo documentation`, `/document-plan`, or `doc plan`.
Optional argument: path to an existing **reviewed** `docs/overview.md` (or equivalent) to seed the plan.
## Outcome
In the **target workspace** (not this toolkit repo unless that is the repo to document):
- `docs/documentation-plan/plan.md` - baby-step documentation plan with progress tracking
- `docs/overview.md` - high-level repo summary (created or refreshed unless user supplied overview)
Product documentation prose follows the **language the user chooses** (pt-BR or English). File paths stay in English.
## Lazy-load
| When | Path |
|------|------|
| Caveman Mode (if active) | `{{TOOLKIT_ROOT}}/skills/_shared/caveman/CAVEMAN.md` - **Lite cap** |
| Plan template, domain checklist | `skills/document-plan/reference.md` or `{{TOOLKIT_ROOT}}/skills/document-plan/reference.md` after sync |
| SDD vs RAG plan boundary | `{{TOOLKIT_ROOT}}/skills/_shared/sdd-artifacts/STORAGE.md` (read section Read-only discovery - do not conflate paths) |
| Context pressure | `{{TOOLKIT_ROOT}}/rules/context-management.mdc` |
## Process
### Step -1b - Caveman Mode (Lite cap)
1. Read `{{SDD_ROOT}}/preferences.json` (create `{ "caveman_mode": false, "caveman_level": "full" }` if missing).
2. If `caveman_mode` is false: continue without compression.
3. If true: load `{{TOOLKIT_ROOT}}/skills/_shared/caveman/CAVEMAN.md`; apply **Lite** participation cap + prefs `caveman_level` (Lite skills never escalate); show once: `[Caveman] Modo ativo (respostas compactas, level={effective}). Digite caveman off para desativar.`
4. Honor `caveman on|off|status|lite|full|ultra` (and `stop caveman` / `normal mode`) during the session.
5. Auto-Clarity + never-compress gates/drafts/paths per `CAVEMAN.md`.
### 0. Workspace and stack
1. Confirm **target repository** (consumer app/service - not the toolkit unless explicit).
2. Detect stack with Glob (see `reference.md` section Stack detection). Summarize: languages, frameworks, layout - **do not** assume a fixed corporate stack.
3. Read `AGENTS.md` / `README.md` if present.
**Not SDD:** this skill writes `docs/documentation-plan/plan.md` only - not `PLAN/PLAN_*.md` or `{{SDD_ROOT}}/<repo-id>/PLAN/`. Feature PRD/PLAN use `sdd-spec` / `sdd-plan` / `sdd-develop` per `STORAGE.md`.
### 1. Documentation language (blocker)
Before creating or updating any file under `docs/` in the target repo, ask once:
> Documentation language for product `docs/` - **pt-BR** or **English**?
Record the choice in the session and in `docs/documentation-plan/plan.md` header (see template).
### 2. Existing artifacts
| Path | Action |
|------|--------|
| `docs/documentation-plan/plan.md` | If exists: read completed steps; resume or ask to overwrite |
| `docs/overview.md` | Use user-provided path, existing file, or generate from exploration |
| Neither | Create both per `reference.md` |
### 3. Explore and draft overview
Glob/Grep/Read: solution layout, main entry points, bounded contexts, external integrations, test layout. Write or update `docs/overview.md` in the chosen language - concise, RAG-friendly, no invented business domain names.
### 4. Write documentation plan
Create or update `docs/documentation-plan/plan.md` using `reference.md` section Plan template:
- Baby steps per business/technical domain
- Steps for integrations, architecture patterns, and folder conventions
- Progress `0/N`, status per step (**Pending** / **Completed**)
- Each step sized for one `document-implement` session where possible
### 5. Context checkpoint
After finishing overview + plan (or after each major planning chunk if the repo is large), follow `context-management.mdc`: at **>= 40%** context, save `plan.md` and pause; ask whether to continue in a new chat.
### 6. Summarize handoff
Report: paths written, language, step count, first pending step id.
```
/document-implement
```
## Must not
- Embed MES, Athena, or organization-specific product context
- Require external wiki APIs, or fixed .NET/Angular versions without detection
- Create `docs/documentation-plan/` inside **this toolkit repo** during porting (only in consumer repos at runtime)
- Write `docs/` before the language question is answered
- Default product doc language without asking
## Handoff
| Situation | Next |
|-----------|------|
| Execute next doc step | `/document-implement` |
| Code change needed | `/developer` or SDD `sdd-develop` |
| Commit docs | `/commit` |
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!