build-docs
Skill114 starsChanged 8 months agoMITView on GitHub
Build documentation packages from configs. Downloads docs, filters with AI, generates SKILL.md. Use when setting up or updating local documentation.
4 other projects use a version of this.
Adopted by
Most recently taken up by
- NathanForster/claude-commandstheir own version, first seen Nov 1, 2026
- NumericalEarth/Breeze.jltheir own version, first seen Oct 17, 2026
- CliMA/Oceananigans.jltheir own version, first seen Oct 17, 2026
- microsoft/semantic-link-labstheir own version, first seen Feb 12, 2026
Other versions4 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 4
- CliMA/Oceananigans.jl★ 1.4ksame name
- microsoft/semantic-link-labs★ 594same name
- NumericalEarth/Breeze.jl★ 90same name
- NathanForster/claude-commands★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository5 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- web/CLAUDE.mdCLAUDE.md
- build-artifactsSkill
- download-docsSkill
The file
We split this skill into the rules it gives an agent
5 rules. 0 are common, 1 is distinctive, and 4 are very specific to this project.
0
Common
Most of the 4 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 25
No args → Process all configs in vault/configs/*.json
- Very specific: line 72
If --force flag was set, pass it to subagents:
- Distinctive: line 83
All subagents run in parallel.
- Very specific: line 138
After install/link, skills are available as /olore-<name>-<version> (e.g., /olore-prisma-latest).
- Very specific: line 149
Use --force on any standalone skill to regenerate even if output exists.
“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Build Docs
Build documentation packages from configs in vault/configs/.
Usage
/build-docs # All configs, all versions
/build-docs prisma # All versions for prisma config
/build-docs prisma@latest # Specific version only
/build-docs prisma zod # Multiple configs
/build-docs prisma --force # Re-download even if existsWorkflow
Phase 1: Parse Arguments
Parse $ARGUMENTS:
- No args → Process all configs in
vault/configs/*.json namewithout@→ All versions for that configname@version→ Specific version only--forceflag → Re-download even if exists
Phase 2: Discover Work
For each config, list versions and check what needs building:
# List configs
ls vault/configs/*.json
# Get versions from a config
jq -r '.versions | keys[]' vault/configs/{name}.json
# Check if version is built (olore-lock.json exists)
test -f vault/packages/{name}/{version}/olore-lock.jsonBuilt = vault/packages/{name}/{version}/olore-lock.json exists
Work = versions without olore-lock.json (unless --force)
Phase 3: Report Plan and Confirm
Show what will be built and ask for confirmation:
Build plan:
prisma@latest - not built
zod@latest - exists (skip)
nextjs@16.1.3 - not built
Will build 2 packages. Proceed?Phase 4: Spawn Subagents (PARALLEL)
CRITICAL: Spawn ALL subagents in a SINGLE message for parallel execution.
For each package that needs building, spawn a package-builder subagent:
Use package-builder agent: prisma@latest
Use package-builder agent: nextjs@16.1.3If --force flag was set, pass it to subagents:
Use package-builder agent: prisma@latest --forceExample with 3 packages: In ONE message, spawn THREE subagents:
package-builder: prisma@latestpackage-builder: zod@latestpackage-builder: nextjs@16.1.3
All subagents run in parallel. Each handles:
- Download docs from source
- Filter files (if needed)
- Generate TOC.md
- Generate SKILL.md
Phase 5: Summary
Collect results from all subagents and report:
Build complete:
✅ prisma@latest - 312 files (126 filtered), tier 3
✅ zod@latest - 17 files, tier 1
⏭️ nextjs@16.1.3 - skipped (exists)
Install with: olore install ./vault/packages/prisma/latestConfig Structure
Config file: vault/configs/prisma.json
{
"name": "prisma",
"_source": { "type": "github", "repo": "prisma/docs", "path": "content" },
"versions": {
"latest": { "ref": "main" }
}
}Package Structure (after build)
vault/packages/{name}/{version}/
├── olore-lock.json # Build metadata
├── SKILL.md # Skill definition (generated)
├── TOC.md # Table of contents (generated)
└── contents/
└── *.md # Documentation filesAfter Build
For development (changes immediately visible):
olore link ./vault/packages/prisma/latestFor testing final package (copies to ~/.olore):
olore install ./vault/packages/prisma/latestAfter install/link, skills are available as /olore-<name>-<version> (e.g., /olore-prisma-latest).
Standalone Skills
Individual phases can be run independently without the full pipeline:
/download-docs prisma@latest # Download + filter only
/build-artifacts prisma@latest # Regenerate SKILL.md + TOC.md onlyUse --force on any standalone skill to regenerate even if output exists.
Dependencies:
/download-docshas no dependencies (runs first)/build-artifactsrequirescontents/from/download-docs
Reference
- Adding new documentation - Config file format, GitHub/URL sources
- Conventions - Config vs lock files, naming rules
- Templates:
vault/packages/docs-packager/1.0.0/templates/(single source of truth)
Requirements
jq- JSON parsinggit- GitHub cloningcurl- URL downloads
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
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.

