By VVeb1250
Five Claude Code skills: graphify-link, codegraph-link, codegraph-affected, mistake-learning, skill-router. Skills auto-load; hooks are registered manually (see README) because the Python launcher differs per OS.
Run only the tests impacted by your changes, using CodeGraph's dependency graph (`codegraph affected`) to map changed source files to the test files that exercise them. Use when the user wants to test a diff/branch fast, run "affected tests only", or avoid a full-suite run after edits. Requires a CodeGraph index (`.codegraph/`).
Wire a project to use CodeGraph MCP (@colbymchenry/codegraph) alongside graphify. Verifies CLI/index/MCP/permissions, auto-runs `codegraph index` if index is missing, writes or removes the managed CLAUDE.md role-split block. Handles unlink when args contain unlink/ยกเลิก/cancel/remove/ลบ/ถอด.
Initialize or connect a project graphify knowledge graph. Use for first-time graph setup, adding a submodule to an existing graph, starting graphify watch, or unlinking a registered root. Not needed every session because reset-graphify.ps1 starts watch automatically when a registered graph exists.
Track and record recurring mistake patterns and provide automation to increment mistake counters in rules/mistakes-index.md via Stop hook.
Local, no-AI skill suggester. A UserPromptSubmit hook scores every installed skill/command against your prompt (TF-IDF cosine) and injects only high-confidence matches, so the model picks the right skill without you remembering names. Multilingual via an optional semantic tier (any language → English skills). Surfaces dormant skills that the harness does not list.
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.
Five Claude Code skills packaged as a plugin.
| Skill | What it does |
|---|---|
graphify-link | Links a project to a graphify knowledge graph. Manages graphify-roots.json, installs fast-path drivers (graphify-build.py, graphify-update.py), starts graphify watch. |
codegraph-link | Wires a project to the CodeGraph MCP (@colbymchenry/codegraph). Writes/removes the managed role-split block in CLAUDE.md. |
codegraph-affected | Runs only the tests impacted by a change. Maps changed source files → impacted test files via codegraph affected, then runs just those (hands off to existing runner skills). Requires a CodeGraph index. |
mistake-learning | Stop hook that reads the session transcript directly (no CLV2 dependency), detects known syntactic mistake patterns, and increments (xN) counters in ~/.claude/rules/mistakes-index.md. Bundles a mistakes-sweep.py health/auto-archive script and seeds the rules/mistakes-*.md files on install. |
skill-router | UserPromptSubmit hook that scores every installed skill/command against your prompt (local TF-IDF cosine, no AI) and injects only high-confidence matches, so the model picks the right skill without you naming it. Surfaces dormant nested skills (e.g. skills/ecc/*) the harness does not list. Silent below threshold → ~0 tokens on unrelated prompts. Index auto-rebuilds when skills are added/removed/edited. |
py launcher (Windows) or python3 (macOS/Linux)graphify CLI — graphify-link skill onlycodegraph CLI — codegraph-link skill only~/.claude/rules/mistakes-*.md — mistake-learning skill auto-seeds these on install if missing (existing files are never overwritten)PowerShell is not required. All hooks are pure Python.
/plugin)Installs the skills natively — no copy step, auto-discovered by the harness.
/plugin marketplace add VVeb1250/WhipForAWeeb-skills
/plugin install whipforaweeb-skills@whipforaweeb-skills
Hooks are not auto-registered by
/plugin. Thepy/python3launcher differs per OS, so the hooks (skill-router, mistake-learning, graphify intercept) must be added to~/.claude/settings.jsonby hand — see Registering hooks below. For a one-shot setup that prints the right snippets for your OS, use the npm installer instead.
npx whipforaweeb-skills
npm install -g whipforaweeb-skills
whipforaweeb-skills
# macOS / Linux
git clone https://github.com/VVeb1250/WhipForAWeeb-skills.git
cd WhipForAWeeb-skills
bash install.sh
# Windows
git clone https://github.com/VVeb1250/WhipForAWeeb-skills.git
cd WhipForAWeeb-skills
pwsh install.ps1
Copy the skill directories you want into ~/.claude/skills/:
# Windows
Copy-Item -Recurse "plugins\skills\graphify-link" "$env:USERPROFILE\.claude\skills\"
Copy-Item -Recurse "plugins\skills\codegraph-link" "$env:USERPROFILE\.claude\skills\"
Copy-Item -Recurse "plugins\skills\codegraph-affected" "$env:USERPROFILE\.claude\skills\"
Copy-Item -Recurse "plugins\skills\mistake-learning" "$env:USERPROFILE\.claude\skills\"
Copy-Item -Recurse "plugins\skills\skill-router" "$env:USERPROFILE\.claude\skills\"
# macOS / Linux
cp -r plugins/skills/graphify-link ~/.claude/skills/
cp -r plugins/skills/codegraph-link ~/.claude/skills/
cp -r plugins/skills/codegraph-affected ~/.claude/skills/
cp -r plugins/skills/mistake-learning ~/.claude/skills/
cp -r plugins/skills/skill-router ~/.claude/skills/
~/.claude/settings.jsonmistake-learning Stop hook — Windows:
{
"hooks": {
"Stop": [
{
"command": "py \"%USERPROFILE%\\.claude\\skills\\mistake-learning\\hooks\\stop-hook.py\"",
"description": "Increment mistake counters on session end"
}
]
}
}
mistake-learning Stop hook — macOS / Linux:
{
"hooks": {
"Stop": [
{
"command": "python3 \"$HOME/.claude/skills/mistake-learning/hooks/stop-hook.py\"",
"description": "Increment mistake counters on session end"
}
]
}
}
mistake-learning sweep (health + auto-archive) — SessionStart, Windows:
{
"hooks": {
"SessionStart": [
{
"command": "py \"%USERPROFILE%\\.claude\\skills\\mistake-learning\\hooks\\mistakes-sweep.py\" --fix-safe --quiet; exit 0",
"description": "Budget check + auto-archive FIXED entries"
}
]
}
}
mistake-learning sweep — SessionStart, macOS / Linux:
npx claudepluginhub vveb1250/whipforaweeb-skills --plugin whipforaweeb-skillsUpstash 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.
Connect to Atlassian products including Jira and Confluence. Search and create issues, access documentation, manage sprints, and integrate your development workflow with Atlassian's collaboration tools.
AI-powered wiki generator for code repositories. Generates comprehensive, Mermaid-rich documentation with dark-mode VitePress sites, onboarding guides, deep research, and source citations. Inspired by OpenDeepWiki and deepwiki-open.
20 modular skills for idiomatic Go — each under 225 lines, backed by 48 reference files, 8 automation scripts (all with --json, --limit, --force), and 4 asset templates. Covers error handling, naming, testing, concurrency, interfaces, generics, documentation, logging, performance, and more. Activates automatically with progressive disclosure and conditional cross-references.
Claude + Obsidian knowledge companion. Sets up a persistent, compounding wiki vault (Karpathy's LLM Wiki pattern). v1.7 "Compound Vault" + v1.8 methodology modes close 5 of 5 priority gaps from the May 2026 compass artifact. Ships: substrate alignment with kepano/obsidian-skills, default Obsidian CLI transport, hybrid retrieval (contextual prefix + BM25 + cosine rerank per Anthropic's Sept 2024 research), per-file advisory locking for multi-writer safety, pre-commit verifier agent, AND methodology modes (LYT / PARA / Zettelkasten / Generic) for first-class organizational support no other Claude+Obsidian competitor offers. v1.7.x audit closure: every BLOCKER + HIGH + MEDIUM + LOW finding from the v1.7.0 audit is CLOSED or DEFERRED-with-rationale. Optional DragonScale Memory extension (log folds, deterministic addresses, semantic tiling lint, boundary-first autoresearch).