skill-authoring plugins

by Mohit Aggarwal

Close the loop between what you keep asking your assistant for and reusable skills: scan your transcripts for recurring patterns, draft a skill in this library's format, test its triggers, and publish a complete release package

by Melodic Software

Doctrine and knowledge playbooks as on-demand skills: boris (Boris Cherny's Claude Code workflow tips, howborisusesclaudecode.com), skill-authoring (Anthropic's internal skill-authoring playbook), fable-5 (Claude Fable 5's operating doctrine, self-authored), and repo-sweep (runs a catalog of hygiene skills through a repository one commit per step). boris and skill-authoring vendor a verbatim upstream baseline; /playbooks:update drift-checks and syncs them (maintainers).

by AIimagined

Skills for a local, tamper-evident evidence ledger of what a coding agent did, claimed, and verified - gates enforced in hooks once wired (run `godmode hooks wire` and Trust the listed commands).

by Grimaldo Stanzani

Superpowers-derived process disciplines in a calibrated register: fit-ranked tool dispatch, capacity dispatch (task-to-model/effort routing on a calibrated rubric), calibration-first skill authoring, red-green TDD, root-cause-first debugging, design-before-code brainstorming, evidence-before-claims verification, technically rigorous review reception, and midweight planned execution (executable-cold plans + per-task subagent review loop) — deduplicated against the craft-collection toolkit and eval-gated from birth.

by jabworks

Codex skills for operating the jabworks/agentic-toolkit repo itself: orientation, skill foundry (create/register/publish), change control, skill-authoring standards, debugging playbook, failure archaeology, plugin schema reference, and research frontier.

by Tamas Mezei

A suite for authoring Claude Code artifacts: reviewing-claude-skills and reviewing-claude-subagents each review their artifact in two passes — structure first, stopping at a structural verdict when the shape fails (fit-for-purpose leads the subagent gate), otherwise a full sweep ending in a computed verdict that honors the artifact's recorded waivers; improving-claude-skills and improving-claude-subagents loop the matching review autonomously — review → apply every blocking finding → commit → re-review — until the verdict is acceptable, the findings plateau, or the round cap is hit, asking only one question: the intent brief at kickoff. The subagent loop ends instead of converting when a review says the subagent should be a skill.

by Coferlandia

Coferlandia Agent Skills and Chat delivery controllers for supervised planning, architecture, software delivery, local/GitHub-native qualification, aggregate releases, emergency hotfix remediation, repository CI adaptation, configuration and release operations, durable project knowledge, developer documentation, evidence-based analysis, and deterministic repository workflows.

by Matheus Malagris

Publish one of your own skills from ~/.claude/skills to this repo: copy it under skills/<name>/, mint its plugin.json, upsert marketplace.json, bump SemVer, and open a PR (CI-gated). Idempotent.

by Webgrip

Author, edit, and audit Claude Code skills for token-efficient, high-trigger-accuracy LLM ingestion: a verified token cost model, description-as-router rules, progressive disclosure into reference files, and skill eval methodology.

by Matheus Malagris

Meta-skill: create new Claude Code skills with proper structure, progressive disclosure (subdirs for reference/workflows/checklists), and bundled resources. Guides the SKILL.md template, description-trigger discipline, and when to split content across files.

by rajool

A toolkit of reusable Claude Code skills and agents: an invoke-only git workflow with guard hooks (no bulk add, no edits on main, no binaries/secrets), deep web research via the Gemini CLI, an Agent Skills authoring/validation kit, a project-agnostic file organizer, a macOS audio meeting recorder plus a processor that transcribes (ElevenLabs Scribe v2, with an automatic Gemini fallback for blocked networks), summarizes, and routes action items into the project's own tools, a document design system (typeset, formerly md-to-pdf) that gives every project its own house style and prints to PDF (a per-project theme of identity, logo, one accent, type pairing, mood and weight set — emphasis is a weight step, never bold — with cover page, letterhead or 16:9 deck slides, running head and folio, a table register with right-aligned tabular figures and ruled totals, financial statements from rows or a QuickBooks export, and first-class RTL/Persian typography — mirrored layout, x-height-matched pairs, a Latin guest face inside Persian sentences, Persian digits, Jalali dates, localized stationery labels), a claim-sourcing helper that links every factual claim to its original independent source, an end-of-day sweep that reconciles every repo with its remote, prunes stale branches, worktrees, and orphaned worktree directories (squash-merge aware, PR-aware), and brings every installed plugin up to its latest published version while flagging manifest/install version drift, a personal end-of-day retro-log engine (five questions, one dated file per day, built for pattern analysis over time), a Chrome browser-automation skill that drives Chrome via the DevTools Protocol (CDP) — snapshot/uid targeting, far more reliable and faster than coordinate-clicking extensions, with named persistent login profiles (isolated throwaway by default, user-defined per-account identities whose sessions survive across runs, guided new-machine onboarding, pinned server version, telemetry off) — a one-command installer that teaches a machine's global CLAUDE.md to render Persian/RTL chat replies correctly (whole-reply RTL widget card, atomic LTR token isolation, English-only plain chat text, BiDi-safe CLI fallback), and an always-on permission guard that blocks destructive commands like rm -rf, and a ship guard that refuses to end a turn while a requested "ship" is still unmerged. Skills act only when invoked.