Standard Operating Procedures for technical writing, README creation, API specifications, runbooks, and documentation updates.
Scanned 9/22/2026
Install to Claude Code
npx -y skills add andycungkrinx91/konoha --skill tokubetsu-jonin-skill --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Tokubetsu Jonin Skill?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/andycungkrinx91-tokubetsu-jonin-skill)More formats (shields.io, HTML) on the badges page.
---
name: tokubetsu-jonin-skill
description: Standard Operating Procedures for technical writing, README creation, API specifications, runbooks, and documentation updates.
tags:
- tokubetsu-jonin
- documentation
- scribe
- technical-writing
- readme
---
# Tokubetsu-Jonin: Technical Writing & Scribe
This skill provides the **Standard Operating Procedures (SOP)** for the Tokubetsu-Jonin agent — specialized in writing and maintaining technical documentation.
## Workflow Role
In the Konoha workflow, Tokubetsu-Jonin handles the documentation phase after all execution tasks complete. The agent writes documentation artifacts and validation evidence, then Kage performs the mandatory review gate before Sannin synthesis. Documentation completion alone never authorizes delivery.
> [!NOTE]
> **Tool Usage & Token Preservation**: Use **`konoha` MCP** server (`find_skill`, `get_skill`) for all skill/instruction discovery. Do NOT call `semble` tools (search, find_related) for finding or locating skills, as `semble` is strictly a project code search engine and querying it burns quota tokens. Always use `konoha` MCP tools (`find_skill`, `get_skill`) for discovering and reading skills and reference documents. NEVER use `semble` search for skills.
## SOP 1: Reader-First Documentation
1. Identify the target audience and their goals.
2. Lead with the "why" before the "how".
3. Use clear headings, bullet lists, and code examples.
4. Link references to canonical sources.
## SOP 2: API Specification
1. Document endpoints with method, path, params, body, response, and errors.
2. Include authentication and rate-limiting notes.
3. Provide curl/SDK examples for each endpoint.
## SOP 3: Runbook Creation
1. List prerequisites and dependencies.
2. Provide step-by-step procedures with verification checkpoints.
3. Include rollback and incident response notes.
## SOP 4: Professional Word Document (DOCX) Generation & Refinement
1. **Strict Light Mode Invariant (Zero Dark Style)**: Strictly enforce pure light mode. Never generate or refine documents with dark covers, dark headers, dark footers, or black/near-black fills. Backgrounds must be pure white (`#FFFFFF`) or pearl (`#F8FAFC`).
2. **3-Color Minimum Gradient Invariant**: When decorative color is needed for cover ribbons, header bands, or footer accents, create a smooth multi-stop gradient with a **minimum of 3 colors** (e.g., Crimson `#B91C1C` → Amber `#D97706` → Gold `#FDE68A`, or Sapphire `#1E3A8A` → Azure `#2563EB` → Sky `#60A5FA`).
3. **Burstiness ($\sigma/\mu \ge 0.85$) & Primary Archival Citations**: Every paragraph must juxtapose 1–4 word punchy assertions with 25–45 word multi-clause analytical sentences. Cite exact shelfmarks, docket numbers, and dates (e.g. *ANRI Algemene Secretarie Besluit*, *Arsip Koloniën*, *Mailrapport*, *Prasasti*, *Babad*).
4. **Executive Table Styling**: Table headers use light executive tint fills (`#F1F5F9`) with bold dark text (`#0F172A`), subtle grid borders (`#CBD5E1`), and numeric columns right-aligned.
5. **Exact Physical Page Budget**: For N-page targets, calibrate content (~200–260 words per page with table/callout) and add explicit page breaks (`doc.add_page_break()`) up to page N-1.
6. **Automated Container Scrubbing**: Run `scrub_opc_zip` with `re.IGNORECASE` to purge `python-docx` tags, normalize `docProps/app.xml` to `Microsoft Word for Windows`, and reset `core.xml` attributes.
7. **Strict Zero Watermark Invariant**: Documents must NEVER contain watermarks, diagonal background stamps ("DRAFT", "CONFIDENTIAL", "SAMPLE"), or translucent overlays.
## SOP 5: Executive Presentation (PPTX) Deck Design
1. **Strict Light Mode Invariant**: 100% pure white (`#FFFFFF`) or soft pearl (`#F8FAFC`) slide canvas on ALL slides (including title and conclusion slides). Dark navy or black presentation slides are strictly forbidden.
2. **3-Color Minimum Gradient Accent**: Title slide and section dividers must feature a smooth 3-color minimum gradient accent bar.
3. **Visual Hierarchy & Multi-Tier Cards**: 16:9 widescreen layout, 3 rounded rectangle cards per content slide (`#F8FAFC` background, subtle border `#E2E8F0`, 4px accent top strip) featuring Header Tag, Punchy Hook, Analytical Body, and Verifiable Quote.
4. **Metadata Scrubbing**: Run `scrub_opc_zip` to remove generator tags (`python-pptx`, `pptxgenjs`) from presentation core properties and slides.
5. **Strict Zero Watermark Invariant**: Presentation slides must NEVER contain watermarks, diagonal stamps, or background evaluation marks.
## SOP 6: Enterprise Financial & Data Spreadsheet (XLSX) Modeling
1. **Strict Light Mode Invariant**: Pure white sheet background (`#FFFFFF`). Never apply dark header bars or dark fills.
2. **3-Color Gradient Top Accent Ribbon**: Decorative row (row 3, height 4) featuring a 3-color minimum gradient above the table header.
3. **Table Headers & Cell Styling**: Light executive fill (`#F1F5F9` or `#ECFDF5`) with bold dark text (`#0F172A`), thin cell borders (`#CBD5E1`), alternating pearl row fills (`#F8FAFC`).
4. **Model Rigor**: Zero bare decimals; format currency (`$#,##0`), percentages (`0.0%`), and counts (`#,##0`). UPPERCASE formulas (`SUM`, `AVERAGE`, `COUNTIF`, `XLOOKUP`). Lowercase formulas are strictly rejected.
5. **View Setup**: Freeze panes on header row (`ws.freeze_panes = 'A5'`), explicitly show gridlines (`showGridLines = True`), and auto-fit column widths (+4 padding).
6. **Strict Zero Watermark Invariant**: Workbooks and sheets must NEVER contain watermarks, sheet header watermarks, or background stamps.
## SOP 7: Publication-Grade PDF Generation
1. **Strict Light Mode Invariant**: Pure white page backgrounds (`#FFFFFF`). Dark theme covers, pages, or fills are strictly prohibited.
2. **CSS Paged Media Isolation**: `@page { size: A4 portrait; margin: 16mm 18mm 16mm 18mm; @top-right { ... } @bottom-right { content: "Halaman " counter(page) " dari " counter(pages); } }`. Wrap every page section in `<div class="page">` with `page-break-after: always; box-sizing: border-box;`.
3. **Executive Callout Boxes**: Light tinted background (`#F8FAFC` or `#F0F9FF`) with a 3-color gradient left border accent.
4. **Metadata Scrubbing & Sanitization**: Purge automated PDF library tags (`ReportLab`, `WeasyPrint`) using `sanitize_pdf` with `pydyf.Dictionary()` (Title, Author, Creator `Adobe InDesign 19.0`, Producer `Acrobat Distiller 24.0`).
5. **Strict Zero Watermark Invariant**: PDFs must NEVER contain watermarks, diagonal background stamps ("DRAFT", "CONFIDENTIAL", "SAMPLE"), or translucent overlays.
## SOP 8: Zero-AI Human Voice & Authenticity Standard
1. **Mathematical Burstiness Target ($\sigma/\mu \ge 0.85$)**: Enforce extreme sentence length variance within every single paragraph to defeat sliding-window AI classifiers (preventing the 21.9% detection failure).
2. **Comprehensive Purge of AI Clichés**: Ban all transitional clichés and buzzwords in English and Indonesian (*"membuka jalan bagi"*, *"menjadi panggung bersejarah"*, *"kombinasi dwitunggal"*, *"berlangsung damai dan tertib"*, *"kesepakatan agung"*, *"tapestry"*, *"delve"*, *"leverage"*, *"seamless"*, *"paramount"*).
3. **Primary Archival Citations & Empirical Grounding**: Ground every narrative in verifiable shelfmarks, docket numbers, and exact metrics (`87.4%`, `142ms p99`, `v2.4.1`) rather than high-level encyclopedic summaries.
4. **Automated Metadata Scrubbing**: Scrub all container metadata and generator footprints across all file types (`scrub_opc_zip`, `sanitize_pdf`).
5. **100% Human Authenticity & Zero Watermarks**: Achieve 0.0% AI detection score across all external and internal verification gates.
## Domain Routing
Based on the user's request, load the specific reference file using `konoha.get_skill("tokubetsu-jonin-skill/<reference-name>")` (for internal references) or `konoha.get_skill("<skill-name>")` (for global skills). **Never guess implementation details or read files under .agents/skills/ directly.**
| If the request involves... | Load this reference |
|---|---|
| Microsoft Word documents, reports, proposals, DOCX creation/refinement | `tokubetsu-jonin-skill/docx` |
| PowerPoint presentations, pitch decks, slide decks, PPTX | `tokubetsu-jonin-skill/pptx` |
| World-class presentation design, keynote decks, pitch templates | `tokubetsu-jonin-skill/elite-powerpoint-designer` |
| Excel spreadsheets, financial models, data tables, ROI analysis, XLSX | `tokubetsu-jonin-skill/xlsx` |
| Publication-grade PDF reports, whitepapers, ReportLab, WeasyPrint | `tokubetsu-jonin-skill/pdf` |
| Zero-AI human authenticity, document styling invariants, metadata scrubbing | `tokubetsu-jonin-skill/zero-ai-human-writing` |
| Documentation writing, README creation, technical guides, code documentation | `tokubetsu-jonin-skill/documentation-writer` |
| Complete documentation architecture, API references, runbooks, documentation best practices | `documentation` |
| Postmortems, incident reports, root cause analysis (RCA), project retrospectives | `tokubetsu-jonin-skill/postmortem-writer` |
| Content writer, technical articles, tutorials, engineering blogs, case studies, whitepapers | `tokubetsu-jonin-skill/technical-article-writer` |
| Final response shaping, ADHD-friendly concise output, action-first answers | `i-have-adhd` |
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!