building-wiki
Skill0 starsChanged 29 days agoMITView on GitHub
Bootstraps and operates a wiki — cited pages compiled from sources by an agent, plus a short `AGENTS.md` schema for one domain. Use when asked to set up a knowledge base or wiki, to write or audit its schema, or to ingest a source into an existing one, answer a question from it, lint it, or refresh it after a source changed.
In this repository13 other files it gives its agents.
- AGENTS.mdAGENTS.md
- writing-skillsSkill
- compressing-knowledgeSkill
- creating-issuesSkill
- designing-cliSkill
- fuckSkill
- functional-domain-modelingSkill
- git-shippingSkill
- no-commentsSkill
- orchestratingSkill
- receiving-code-reviewSkill
- tracking-tasksSkill
- writing-japaneseSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 0 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 28
Read the repository's AGENTS.md, then run the operation — ingest, query, lint, or update — as references/operations.md defines it.
- Very specific: line 32
Follow references/bootstrapping.md: interview, lay out the bundle, write the schema, pilot.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Building a Wiki
This is the LLM Wiki pattern, instantiated for one repository's domain.
Three layers: immutable sources the human picks, pages the agent writes and maintains, and a schema section in the repository's root AGENTS.md holding only what is specific to this domain. The human sources, explores, and asks; the agent does the bookkeeping — cross-references, freshness, contradictions.
The format is fixed; the amount of structure is not. Start with the least that answers the questions asked so far; add a page, a type, or a schema line only when its absence caused a wrong answer or a repeated lookup.
Sources sit upstream of the implementation: specifications, decisions, vendor documents, meeting notes. The repository's own code and pull requests are what the wiki informs, never a source — even when the wiki arrives after the code, a page states what the code should do, and a disagreement is a finding about the code.
When Not To Use
- Retrieval over documents as they stand — RAG and full-text search complement a wiki rather than replace one.
- A single source summarized once, with nothing accumulating.
- Knowledge already held in a machine-readable schema (OpenAPI, Protobuf) — reference it instead.
Operating an Existing Wiki
Read the repository's AGENTS.md, then run the operation — ingest, query, lint, or update — as references/operations.md defines it.
Bootstrapping a Wiki
Follow references/bootstrapping.md: interview, lay out the bundle, write the schema, pilot.
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.