Use when writing or reviewing pynchy documentation, deciding where to document things, updating the docs, checking doc consistency, or fixing broken links. Covers information architecture, writing philosophy, tree-shaped navigation, doc-code coupling, no hard-coded usernames, extensibility framing for pluggable subsystems, and when to add code comments.
Scanned 5/27/2026
Install via CLI
openskills install crypdick/pynchy---
name: Documentation Manager
description: Use when writing or reviewing pynchy documentation, deciding where to document things, updating the docs, checking doc consistency, or fixing broken links. Covers information architecture, writing philosophy, tree-shaped navigation, doc-code coupling, no hard-coded usernames, extensibility framing for pluggable subsystems, and when to add code comments.
---
# Documentation Manager
Helps decide where to document things and maintain consistency across Pynchy docs.
## Style Guide
- Write for a **user trying to achieve a goal**. Don't chronicle the evolution of the codebase. Don't go into unnecessary technical details that are not relevant to the user's goal.
- Follow the [Google Style Guide](https://developers.google.com/style). Write in the present tense. Don't use "we" or "they". Write in the active voice. Don't use "is" or "are".
- Read `docs/contributing/contributing-docs.md` for the full documentation philosophy and information architecture rules.
## Where to Document What
Quick decision tree:
**New feature?**
- Architecture decision → `docs/architecture/` (find the relevant topic file, or create a new one)
- Installation requirement → `docs/install.md`
- Security implication → `docs/architecture/security.md`
- Development workflow change → `.claude/skills/pynchy-dev/SKILL.md`
**Bug fix?**
- If it needs install change → `docs/install.md`
- Usually: No doc update needed
**Refactoring?**
- If user-visible → Update relevant docs
- If internal only → No doc update
## File Purposes
| File | What Goes There |
|------|-----------------|
| `README.md` | Philosophy, quick start, high-level overview |
| `docs/install.md` | Complete installation guide |
| `docs/architecture/security.md` | Security model, threat analysis |
| `docs/architecture/index.md` | Architecture overview and links to topic pages |
| `docs/architecture/*.md` | One topic per file (containers, routing, tasks, etc.) |
| `.claude/skills/` | Development context skills for Claude Code agents |
## Information Architecture Rules
1. **Single source of truth** — Every concept explained in exactly one place. Cross-link, don't duplicate.
2. **Tree-shaped navigation** — Root files have links + short summaries. Leaf files have the actual content.
3. **Small, focused files** — One topic per file. If it covers multiple concerns, split it.
## Validation
**Before committing:**
```bash
# Check for broken links
uv run mkdocs build --strict
```
**After moving/renaming files:**
1. Search for all references to old name
2. Update each reference
3. Update `mkdocs.yml` nav
4. Test with `uv run mkdocs build --strict`
## Common Mistakes
- Duplicating content across files (link instead)
- Chronological explanations ("First we tried X...")
- Mixing audiences (keep README brief, details in docs/)
- Forgetting to update references after renames
- Writing mega-files that cover multiple topics
## Link Checking
Link validation runs automatically in pre-commit hooks. If docs have broken links, the commit will fail.
To manually check: `uv run mkdocs build --strict`
No comments yet. Be the first to comment!