Stats
Actions
Available In
Tags
Plugin
efficiency-audit
Analyzes recent Claude Code transcripts to surface recurring inefficiencies.
README
claude-marketplace
A collection of Claude Code plugins — installable via the litianningdatadog-marketplace plugin marketplace. Each plugin delivers a skill that extends Claude's capabilities with a specialized, repeatable workflow.
Each plugin has a .claude-plugin/plugin.json manifest and a skills/<name>/SKILL.md that becomes Claude's operating instructions once the skill activates. Supporting scripts live under <plugin>/scripts/. This README is the entry point for humans browsing or installing the repo.
Plugins
| Skill | What it does |
|---|---|
efficiency-audit | Analyzes recent Claude Code transcripts to surface recurring friction — corrections, re-explained context, tool-call failures, failing hooks — grouped by recurrence count and dominant project. LLM synthesis drafts ranked CLAUDE.md rules with token-savings estimates; idempotent marker blocks prevent duplicate accumulation across re-runs. |
hook-doctor | Inspects and repairs installed plugin hook configurations (hooks.json). Detects unquoted ${CLAUDE_PLUGIN_ROOT} commands that fail in agent-mode, reports them, and applies safe idempotent fixes with explicit opt-in. |
quicknotes | Low-friction quick-note capture and management. Centralized markdown notes (~/.quicknotes) with date/project/dir metadata, tags, fuzzy search, references, and time/location reminders; completing a note deletes it. Capture via the qn CLI (instant), /qn, or natural language. |
Installing plugins
This repo is a Claude Code plugin marketplace. Add it once, then install skills à la carte:
# Add the marketplace (once per machine)
/plugin marketplace add litianningdatadog/claude-marketplace
# Install skills
/plugin install efficiency-audit@litianningdatadog-marketplace
/plugin install hook-doctor@litianningdatadog-marketplace
/plugin install quicknotes@litianningdatadog-marketplace
Plugins auto-update when you run /plugin marketplace update. Each skill activates
automatically when your request matches the triggers in its SKILL.md description.
Development
No build system or third-party dependencies — skills are Markdown plus scripts. Scripts
that ship tests use the Python standard-library unittest, run from the script's
directory. Per-skill setup, CLI usage, and test commands live in each skill's README.
Contributing a new skill
- Create a top-level directory named in kebab-case (matches the skill's
name). - Add a
.claude-plugin/plugin.jsonwithnameanddescription. - Add
skills/<name>/SKILL.mdas the agent procedure. Front-load trigger phrases in the frontmatterdescription— it's the only text read when deciding whether to activate. Reference supporting scripts via${PLUGIN_ROOT}/scripts/...(resolved dynamically at runtime). - Put supporting code under
<skill>/scripts/(stays at the plugin root). - Add the plugin entry to
.claude-plugin/marketplace.jsonand the skill to the table above.
See CLAUDE.md for the conventions Claude follows when working in this repo.
Stats
Version1.0.0
LanguagePython
Stars0
MaintenanceExcellent
Last Commit
Added
Available In
Tags
Categories
