local-wiki
Skill1 starsChanged 35 days agoMITView on GitHub
Build or update a source-backed local repository wiki, validate its navigation, or export a linked Markdown bundle and optional offline static site. Use for implemented-system documentation, not future design or automatic documentation cleanup.
- Installs packages
1 other project uses a version of this.
Adopted by
Most recently taken up by
- rcdelfin/agentkittheir own version, first seen Aug 28, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- rcdelfin/agentkit★ 2same 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 repository27 other files it gives its agents.
- architectSkill
- audit-discoverySkill
- audit-initSkill
- audit-nextSkill
- audit-resolutionSkill
- audit-statusSkill
- audit-triageSkill
- audit-verificationSkill
- code-to-prdSkill
- codex-ubuntu-serverSkill
- copilot-review-customizerSkill
- create-pluginSkill
- debugSkill
- deep-reviewSkill
- design-promptsSkill
- diagnose-sessionSkill
- feature-scatteringSkill
- human-execution-runbookSkill
- humanizerSkill
- init-repo-instructionsSkill
- inspect-pluginSkill
- install-agent-pluginsSkill
- recoverSkill
- review-feedbackSkill
- review-repo-instructionsSkill
- uncompromising-handoverSkill
- update-pluginSkill
The file
We split this skill into the rules it gives an agent
5 rules. About 0 are common, 3 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
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
- Distinctive: line 18
Read repository intent, scoped instructions, manifests, real entry points, commands, and existing docs.
- Distinctive: line 28
Write each chosen page with a title, short summary, source-relative evidence in code spans, and links to related wiki pages.
- Very specific: line 39
Use lowercase kebab-case filenames for newly generated pages and keep individual pages focused.
- Very specific: line 68
Search needs HTTP access to the site.
- Distinctive: line 91
Do not generate a companion report, automatic statistics, extra archive, or publication workflow.
≈ “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
Local Wiki
Scope and Modes
Choose inspect/validate, author/update Markdown, export Markdown, or build/serve HTML from the request. Inspection and plain validation are read-only. Creating a wiki does not authorize a site, bundle, local server, or changes to existing project documentation. Keep generated output separate from user-maintained documentation; preserve existing narrative on explicit updates.
Source-backed Authoring
- Read repository intent, scoped instructions, manifests, real entry points, commands, and existing docs. Trace representative handlers, persistence, events/jobs, state, and module boundaries instead of inferring them from names.
- Reconcile discovered capabilities with the relevant directory/package list. State meaningful coverage gaps and unknowns. Do not create a survey artifact or spawn a fleet of agents; the ordinary workflow is sequential.
- Choose a proportional table of contents. A small project may need only an overview and one workflow page. Add subsystem, API, operations, glossary, or contributor pages only when useful. History, statistics, and extensive foundation pages are optional, never required sections.
- Write each chosen page with a title, short summary, source-relative evidence in code spans, and links to related wiki pages. Distinguish observed source, runtime evidence, inferred explanation, and unresolved behavior. Avoid claims of exhaustive coverage or operational success from source inspection alone.
- In a dedicated generated wiki directory, finalize navigation and metadata, then validate. Finalization replaces only marked index blocks and generated metadata, adding missing indexes; it preserves prose outside those blocks. It is a write operation, not part of an inspection-only request.
- Inspect unresolved links, headings, and important source claims. Report actual helper results and limitations; do not mark inferred behavior as reproduced.
Use lowercase kebab-case filenames for newly generated pages and keep individual
pages focused. Split unwieldy pages when useful, not to fill a template. Existing
.wiki-meta.json page order is respected; otherwise indexes precede sorted pages.
The metadata is consumed by navigation and validation, not migration bookkeeping.
Installed Helpers
Python 3.10+ is required. Resolve SKILL_DIR to the directory containing this
loaded skill file. It is a shell variable set from the actual installed resource
path, not a universal host-provided variable. Helpers and HTML templates resolve
relative to that installed location; no checkout or editable install is needed.
The following are independent operations; run only those authorized:
python3 -B "$SKILL_DIR/scripts/local_wiki.py" finalize /path/to/generated-wiki --repo /path/to/repo
python3 -B "$SKILL_DIR/scripts/local_wiki.py" validate /path/to/generated-wiki
python3 -B "$SKILL_DIR/scripts/local_wiki.py" build-markdown /path/to/generated-wiki --out /path/to/wiki-bundle.md
python3 -m pip install -r "$SKILL_DIR/scripts/requirements.txt"
python3 -B "$SKILL_DIR/scripts/local_wiki.py" build-site /path/to/generated-wiki --out /path/to/wiki-site
python3 -B "$SKILL_DIR/scripts/local_wiki.py" serve /path/to/wiki-site --port 8765finalize, validate, and build-markdown use only the standard library.
Installing dependencies is needed only for requested HTML rendering and requires
permission to change the selected Python environment. The optional renderer uses
Jinja2 and markdown-it-py; it does not fetch browser scripts or invoke formatters.
Mermaid remains readable fenced source offline rather than a remotely rendered
diagram. The site contains local navigation, search JSON, CSS, and JavaScript.
Search needs HTTP access to the site; the explicit serve operation runs in the
foreground until interrupted. A generated file is not evidence of a live server.
Output and Validation Boundaries
- The Markdown bundle must be a new file outside the wiki; existing output is refused. Preserve page anchors and relative source evidence. Attachment links are rebased to the original wiki files; move those files with the bundle when sharing it, or explicitly choose a different asset delivery strategy.
- Site output must be outside and not an ancestor of the source wiki. An existing
directory is replaceable only when marked by this renderer (
.droid-wiki-site). Its basename never authorizes deletion. Never add a marker to unrelated content to bypass refusal; choose a new destination. Marked site directories are wholly generated, so do not store manual edits there. - Plain validation checks structure, metadata consistency, and inline local-link
file targets. It is not a full Markdown dialect/fragment validator or proof of
the prose. Review heading fragments, reference-style links, images, and browser
navigation separately. The optional
validate --auto-fixrewrites headings and some links; run it only when that edit is explicitly authorized. - A nonzero helper result is a failure, not a completed wiki. Preserve partial output for inspection; do not delete source or unrelated directories to retry.
Return the requested output paths, actual commands/results, and remaining gaps. Do not generate a companion report, automatic statistics, extra archive, or publication workflow.
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.

