agentleFS
mgonzalezbailemgonzalezbaile/personal-assistant/.claude/skills/wiki/SKILL.md

wiki

Skill8 starsChanged 6 months agoMITView on GitHub

Use when the user wants to ingest sources into the wiki, query existing knowledge, or lint the knowledge base for issues. Manages a persistent, interlinked knowledge base (LLM Wiki) with three operations: ingest, query, lint.

43 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use ai-slop-cleaner, ralplan and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • Jurel89Skill

    ai-slop-cleaner

    Clean AI-generated code slop with a regression-safe, deletion-first workflow and optional reviewer-only mode

    Also in 4 projects that use this

  • ibahgatSkill

    ralplan

    Create a strict, quality-first execution plan with explicit gates and rollback-ready sequencing.

    Also in 4 projects that use this

  • Yeachan-HeoSkill

    ultragoal

    Create and execute durable repo-native multi-goal plans over Codex goal mode artifacts.

    Also in 3 projects that use this

  • woosikkimSkill

    deepinit

    Generate hierarchical AGENTS.md codebase maps for agent context

    Also in 3 projects that use this

  • ibahgatSkill

    hud

    Control omi visual HUD profile and keep status output readable in normal, compact, or hidden mode.

    Also in 3 projects that use this

  • klauswgSkill

    deep-interview

    在明确批准执行之前,通过数学化歧义门控进行苏格拉底式深度访谈

    Also in 4 projects that use this

The file

Show its rules and text

We split this skill into the rules it gives an agent

13 rules. 0 are common, 5 are distinctive, and 8 are very specific to this project.

  • 0

    Common

    Most of the 37 edited versions we read keep these.

  • 5

    Distinctive

    This author's own choices. Worth a look.

  • 8

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read the source document fully.

    line 29
  2. Very specific:

    Update existing project and concept pages with new information from the source.

    line 44
  3. Distinctive:

    Add new pages to the appropriate section (Projects, Concepts, Preferences)

    line 46
  4. Distinctive:

    Update one-line summaries if they've become stale

    line 47
  5. Very specific:

    Delete the source file from sources/ — it has been fully processed.

    line 48
  6. Very specific:

    Never modify files in sources/.

    line 51
  7. Very specific:

    Every new page must be added to memory/CLAUDE.md.

    line 54
  8. Very specific:

    Delete each source file after processing — sources/ should be empty when all ingestion is done.

    line 55
  9. Very specific:

    Use relative links between wiki pages (they're all siblings in memory/).

    line 56
  10. Very specific:

    Read memory/CLAUDE.md to identify relevant pages.

    line 65
  11. Very specific:

    If the answer represents a valuable synthesis (comparison, analysis, connection), offer to file it as a new wiki page of type synthesis.

    line 68
  12. Distinctive:

    Always cite which wiki pages informed the answer.

    line 71
  13. Distinctive:

    If the wiki doesn't contain enough information, say so and suggest sources to ingest.

    line 72

“Common” is measured: kept by at least 6 in 10 of the 37 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

LLM Wiki

A persistent, compounding knowledge base. The LLM writes and maintains the wiki; {{NAME}} curates sources, directs analysis, and asks questions.

Architecture

Three layers:

  • Sources (sources/) — raw, immutable input documents (transcripts, articles, PDFs). Never modified by the LLM.
  • Wiki (memory/) — flat directory of LLM-generated markdown pages. All pages are siblings. memory/CLAUDE.md is the navigation hub.
  • Schema (CLAUDE.md + this skill) — governs structure and operations.

Operations

Ingest

Process a raw source from sources/ into wiki pages. Invoked via the /project:wiki-ingest command.

Input: path to a source file.

Steps:

  1. Read the source document fully.
  2. Identify key information:
    • Decisions made and their rationale
    • Action items and who owns them
    • People mentioned (names, roles, context) — capture on the relevant project/concept page, never as standalone pages
    • Concepts/themes discussed
    • Contradictions with existing wiki content
  3. For each new concept/theme that deserves its own page, create one:
    ---
    type: concept
    name: Concept Name
    first_seen: YYYY-MM-DD
    ---
  4. Update existing project and concept pages with new information from the source. Add cross-references ([link text](other-page.md)) between related pages.
  5. Update memory/CLAUDE.md:
    • Add new pages to the appropriate section (Projects, Concepts, Preferences)
    • Update one-line summaries if they've become stale
  6. Delete the source file from sources/ — it has been fully processed.

Rules:

  • Never modify files in sources/. They are immutable.
  • Never create person pages. People are documented inline on the project or concept pages where they are relevant (e.g., a collaborators table or section).
  • One page per entity (project, concept). If a page already exists, update it — don't create a duplicate.
  • Every new page must be added to memory/CLAUDE.md.
  • Delete each source file after processing — sources/ should be empty when all ingestion is done.
  • Use relative links between wiki pages (they're all siblings in memory/).
  • Frontmatter type field is required on all pages: project, concept, preference, or synthesis.

Query

Answer questions using the wiki as the primary knowledge source.

Steps:

  1. Read memory/CLAUDE.md to identify relevant pages.
  2. Read those pages.
  3. Synthesize an answer with citations to specific wiki pages.
  4. If the answer represents a valuable synthesis (comparison, analysis, connection), offer to file it as a new wiki page of type synthesis.

Rules:

  • Always cite which wiki pages informed the answer.
  • If the wiki doesn't contain enough information, say so and suggest sources to ingest.

Lint

Health-check the wiki. Run periodically or on request.

Checks:

  1. Orphan pages — files in memory/ not listed in memory/CLAUDE.md.
  2. Broken links — references to pages that don't exist.
  3. Stale content — pages not updated in 30+ days that reference active projects.
  4. Missing cross-references — pages that mention entities with their own page but don't link to them.
  5. Contradictions — claims in one page that conflict with another.
  6. Missing pages — entities mentioned across multiple pages that lack their own dedicated page.
  7. Index drift — one-line summaries in memory/CLAUDE.md that no longer match page content.

Output: a report listing issues found and suggested fixes. Apply fixes only with confirmation.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.