By jim-bo
Build and maintain a cited, cross-linked knowledge wiki from a corpus of source documents. Domain-agnostic: a per-project config defines the ontology, page templates, and tooling. Ships specialized agents (compile, entity-page, crosslink, synthesize, query, lint, claims) plus a setup wizard and orchestrator.
(Opt-in) Read one compiled source page and emit a sidecar JSON of machine-verifiable structured claims, per the claim kinds defined in config. Idempotent. Use only when the project config enables claims and the domain has data- or API-verifiable quantitative assertions. <example> Context: A config with claims.enabled is in use; user wants verifiable claims. user: "Extract checkable claims from source 26901067." assistant: "I'll dispatch claims-extractor for that id." <commentary>Structured, verifiable claim extraction is this agent.</commentary> </example>
Scan an explicit list of wiki pages and rewrite bare mentions of known entities as single-bracket markdown links to their canonical pages. Read-only on raw sources, edit-only on the wiki. Use after new source or entity pages are written. <example> Context: A new source page and several entity pages were just written. user: "Crosslink the pages we just touched." assistant: "I'll dispatch crosslinker over that explicit path list." <commentary>Linking bare mentions across a bounded page set is this agent.</commentary> </example>
Create or update ONE wiki entity page from the source documents that cite it. Idempotent — merges into an existing page rather than overwriting. Use after source-compiler returns entity lists for newly compiled sources. <example> Context: source-compiler returned that sources A and B both mention gene EGFR. user: "Update the EGFR entity page with sources A and B." assistant: "I'll dispatch entity-page-writer for kind=gene id=EGFR." <commentary>One entity, merged from its citing sources — this agent.</commentary> </example>
Compile ONE raw source document into a structured wiki source page and return the entities it found so the orchestrator can fan out to entity-page-writer. Use when ingesting a new document into a knowledgizer knowledge base. <example> Context: A new raw document has been added to the corpus. user: "Compile source 12345678 into the wiki." assistant: "I'll dispatch the source-compiler agent for id 12345678." <commentary>Single-document ingest is exactly this agent's job.</commentary> </example>
Read across source pages for a given theme and write or update a wiki theme page, surfacing agreements and contradictions. Use when the user wants a cross-document synthesis rather than a single-source page. <example> Context: The corpus has many sources touching one topic. user: "Synthesize what the corpus says about EGFR resistance." assistant: "I'll dispatch theme-synthesizer with that theme and a slug." <commentary>Cross-source synthesis with conflict-surfacing is this agent.</commentary> </example>
Stand up a new knowledgizer knowledge base by interactively building its domain config and scaffolding. Use when the user says "set up knowledgizer", "start a new knowledge base", "configure knowledgizer", "create the domain config", or is in a repo that has no knowledgizer.config.yaml yet.
Orchestrate building and maintaining a cited, cross-linked knowledge wiki from source documents. Use when the user wants to "add a source/paper/document", "ingest a batch", "rebuild the wiki", "synthesize a theme", "ask the knowledge base", or "reprocess after a template/ontology change" in a repo that has a knowledgizer.config.yaml. Dispatches the source-compiler, entity-page-writer, crosslinker, theme-synthesizer, wiki-querier, wiki-linter, and claims-extractor agents.
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Build and maintain a cited, cross-linked knowledge wiki from a corpus of source documents — papers, cases, contracts, transcripts, internal docs. One config file makes the same agents work for any domain.
Knowledgizer generalizes the cbio-kb
cancer-genomics workflow into a domain-agnostic, installable plugin. It is both
a plugin and a marketplace, so a single git repo is your one place to
update everything.
/plugin marketplace add jim-bo/knowledgizer
/plugin install knowledgizer@knowledgizer
Then in any project run the setup skill once to create the domain config, and use the orchestrator skill to ingest documents.
Note: plugins cannot bundle executables. If your project uses a token-efficient wiki/index CLI (like cbio-kb's
cbio-kb), install it separately and point to it in the config'stoolingblock. Without one, the agents fall back to narrow file reads.
Skills
| Skill | Use it to |
|---|---|
knowledgizer-setup | Interactively build knowledgizer.config.yaml and scaffold the wiki/templates for a new domain. The "start-up step." |
knowledgizer | Orchestrate the loops: add a source, add a batch, synthesize a theme, query the KB, reprocess after changes. |
Agents (dispatched by the orchestrator, run in isolated context for cheap parallel fan-out)
| Agent | Job |
|---|---|
source-compiler | Compile one raw document into a structured page; return its entities. |
entity-page-writer | Create/merge one entity page from its citing sources (idempotent). |
crosslinker | Rewrite bare entity mentions as markdown links to canonical pages. |
theme-synthesizer | Synthesize a cross-source theme; record agreements and conflicts. |
wiki-querier | Answer a question against the wiki; file a cited answer. |
wiki-linter | Audit structure (read-only). |
claims-extractor | (Opt-in) Emit machine-verifiable structured claims as sidecar JSON. |
Reference
reference/config-schema.md — every field of knowledgizer.config.yaml.reference/concepts.md — the methodology and invariants.Example
examples/cbio-kb/ — the cancer-genomics instantiation, mapping each original
cbio-kb concept onto a config field.Each knowledge base has one knowledgizer.config.yaml at its repo root defining
the entity kinds, page templates, source sections, ontology rules, and optional
tooling. The generic agents read it first and carry no hardcoded domain
assumptions — add an entity kind to the config and the whole pipeline extends
with no code change. See reference/config-schema.md.
v0.1.0 — foundation. Expected to iterate, especially on how far the domain-specific bits (claim schemas, ontology discipline) push into config versus per-project hard-rules files.
MIT
npx claudepluginhub jim-bo/knowledgizer --plugin knowledgizerUpstash Context7 MCP server for up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your LLM context.
Build and maintain an LLM-curated personal knowledge base in your project — Andrej Karpathy's LLM Wiki pattern, designed to scale to thousands of pages without becoming a context bottleneck. Now with an optional compiled graph layer for typed, provenance-backed relationships.
Complete creative writing suite with 10 specialized agents covering the full writing process: research gathering, character development, story architecture, world-building, dialogue coaching, editing/review, outlining, content strategy, believability auditing, and prose style/voice analysis. Includes genre-specific guides, templates, and quality checklists.
Comprehensive startup business analysis with market sizing (TAM/SAM/SOM), financial modeling, team planning, and strategic research
v9.44.1 — Patch release for Gemini environment/version detection and qwen auth gating. Run /octo:setup.
Comprehensive .NET development skills for modern C#, ASP.NET, MAUI, Blazor, Aspire, EF Core, Native AOT, testing, security, performance optimization, CI/CD, and cloud-native applications