claudecodex

llm-wiki

Generate a navigable, version-controlled multi-page Markdown wiki from a codebase, engineered for on-demand retrieval as Claude's context while staying human-readable. Two-pass outline-then-fill generation across a parallel agent swarm of planner, diagram author, page writers, and index synthesizer; consumes a code-atlas index as ground truth when present, else self-scans. Pure static Markdown (no embeddings), Mermaid diagrams derived from the dependency graph, per-page source provenance, and git-blob hash-diff staleness detection that regenerates only stale pages. A SessionStart hook loads the wiki index at session start

0 on GitHubby MisterVitoProv0.1.1MITupdated 1 month agodevelopment
/

Plugin llm-wiki from marketplace esper in MisterVitoPro/esper. Commands read from the repo's manifest — 1 month ago.

This plugin runs code on your machine — its hooks fire automatically on events like tool use. The exact commands are listed below; read them before installing.

What it installs

Skills (2)

  • llm-wiki:generate

    Generate a navigable, multi-page Markdown wiki for this codebase under .llm-wiki/. Builds a shared substrate once (reusing a code-atlas index when present, else self-scanning), plans a file-disjoint page set, fans out writer agents in parallel waves to author one page each, derives Mermaid diagrams

  • llm-wiki:update

    Incrementally refresh the wiki under .llm-wiki/. Diffs current file hashes against the state cache (git-blob OIDs, resilient to rebases and branch switches), maps changed source files to the pages they feed via the provenance ledger, and regenerates only stale pages -- or does a full rebuild when ch

Subagents (6)

  • wiki-diagram-author

    llm-wiki agent that produces validated Mermaid diagram blocks for the wiki: a component diagram and dependency diagram derived from the dependency/import graph (deterministic, not invented), plus key sequence/data-flow diagrams. Returns diagram blocks keyed by the target page_id so writers embed the

  • wiki-index-synthesizer

    llm-wiki agent that runs after all writers. Reads every page's returned status JSON and frontmatter, builds the navigation index (.llm-wiki/index.md), computes backlinks, validates cross-links flagging any that dangle, and returns the index plus a pages provenance map for the state cache. Derives th

  • wiki-module-writer

    llm-wiki agent that writes one module, concept, reference, or data-model page. The N-way parallel workhorse: each invocation owns exactly one output Markdown file and receives that page's title, section outline, scoped source contents, related slugs, and any Mermaid blocks to embed. Authors its own

  • wiki-overview-writer

    llm-wiki agent that writes the top-level spine pages (home, overview, getting-started, architecture, data-flow, tech-stack, glossary). Generic template parameterized per invocation with one page's title, type, section outline, scoped source contents, and any Mermaid diagram blocks to embed. Writes e

  • wiki-planner

    llm-wiki agent that plans the wiki. Reads the pre-computed substrate (file index, module boundaries, tech stack, and a code-atlas graph when present) and returns a file-disjoint page plan: a spine of guaranteed pages plus discovered module and concept pages, grouped into waves of <=6 writers ordered

  • wiki-reviewer

    Optional llm-wiki agent that adversarially verifies one already-written page against the source files it claims to document. Re-reads the page and its source_files and flags claims the source does not support (hallucinations, overstatements, stale assertions), plus material facts the page omits. Ret

Hooks (1)

  • SessionStart

    node "${CLAUDE_PLUGIN_ROOT}/scripts/session-start.js"

View source on GitHub

More from MisterVitoPro/esper

All 7 plugins →

qa-swarm

claudecodex

Claude Code and Codex quality pipeline: deploy parallel security, correctness, performance, architecture, data-flow, and concurrency reviewers; rank and corroborate evidence-backed findings; generate a fix spec and test plan; then implement selected fixes through a test-first loop.

2 skills · 16 agents

0

plan-runner

claudecodex

Execute Markdown plans as dependency-ready task DAGs: every task in its own Git worktree, independently verified before it integrates, with durable recovery evidence and human-reviewed PR delivery. Requires Git. Works in Claude Code and Codex.

2 skills · 6 agents · hooks

0

code-atlas

claudecodex

Claude Code and Codex architecture index generator: scan a codebase once, load the curated map at session start, refresh it incrementally, and query its semantic dependency graph deterministically for dependencies, impact analysis, blast radius, and risk filters; the map skill upserts a marker-delimited usage block into CLAUDE.md/AGENTS.md so later sessions consult the atlas

3 skills · 4 agents · hooks

0

migration-runner

claudecodex

Vulnerability-aware dependency upgrade orchestrator across 7 ecosystems (npm, Python, Go, Rust, Java, Kotlin, C#); detects outdated packages, queries OSV.dev for CVEs, recommends the safest-yet-most-recent version per package, then executes wave-by-wave with build/typecheck/test verification and clean git rollback on failure

2 skills · 4 agents · hooks

0

ideas

claudecodex

Claude Code and Codex idea-to-build pipeline: turn a raw idea into an audited design spec through a token-conscious interview with scope-sized batched questions, a category-coverage elicitation floor, an existing-system baseline, and a resumable three-status ledger. A binding ledger audit and biggest-miss critic gate the EARS-based spec. After approval, the Ideas plan skill emits a plan-runner-ready flat task list with full criteria plus a script-generated bundle: a planned source skeleton, a versioned skeleton manifest, and a typed dependency graph with a self-contained per-task Graph context. A completion gate offers a /clear-then-paste plan-runner handoff (fresh context for the execution engine), inline or subagent execution, GitHub tickets, or stop. The Ideas tickets skill projects ready tasks to a parent issue plus linked sub-issues using the gh CLI. Specs, plans, and ADRs land under the project's existing docs root, auto-detected. Not for typos, renames, or one-line fixes.

3 skills · 2 agents

0

jupiter

claudecodex

Consolidate scattered specs into a canonical master-spec tree with a navigable index, and append stubs for undocumented public surface found in code

2 skills · 2 agents

0

Similar plugins

superpowers

claudecodex

Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques

15 skills · hooks

obra/superpowers 316,836

context7

claudecodex

Up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your coding agent's context.

1 MCP server

upstash/context7 64,709

Build powerful, stateful CLI interfaces for any GUI application using the cli-anything harness methodology.

5 commands

HKUDS/CLI-Anything 58,665

Automates browser interactions for web testing, form filling, screenshots, and data extraction

12 skills

vercel-labs/agent-browser 50,373

Use Alibaba Open Code Review from local Codex.

2 skills

alibaba/open-code-review 50,068

slidev

claude

Claude Code skill for creating and presenting Slidev slidedecks

1 skill

slidevjs/slidev 49,452

Maintain llm-wiki?

This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.

claude & codex plugins7 plugins