A collection of Claude Code plugins providing engineering standards for AI coding agents. Plugins are organized around use cases — load the ones that match what you are building.
Guides: docs/TESTING_WORKFLOW.md — how ceh-testing and the three stack
testing skills route between each other, with the trigger phrases and sequence for each moment.
You should not have to remember 22 plugins. Pick the bundle that matches the situation you are in; its dependencies install and enable automatically.
| Bundle | Install when |
|---|---|
ceh-scenario-service-greenfield |
Starting a Python backend service from nothing |
ceh-scenario-service-iterate |
Working on a Python backend service that already ships |
ceh-scenario-library-greenfield |
Starting a Python library from nothing |
ceh-scenario-library-iterate |
Working on a Python library that already ships |
ceh-scenario-webapp-greenfield |
Starting a web frontend from nothing |
ceh-scenario-webapp-iterate |
Working on a web frontend that already ships |
ceh-scenario-editorial |
Writing what readers see — blog posts, user/operator docs, discoverability |
/plugin install ceh-scenario-service-iterate@ceh-plugins --scope user
A -greenfield bundle contains everything its -iterate twin does, plus scaffolding and business
planning — so the phase transition is a no-op. You never switch bundles; you just stop reaching for
the planning skills.
Several plugins are deliberately outside every bundle: ceh-fabled, ceh-advisor and
ceh-orchestration (experimental),
and ceh-ops, ceh-summarize-chat, ceh-lessons-learned (installed once, on
their own, when you want them). A dependency is always installed — defaultEnabled: false does not
protect one — so omission is the only way to keep them opt-in.
The table below is the reference list of what those bundles are made of.
| Plugin | Install as | Contents |
|---|---|---|
| Agent Coding Contract | ceh-coding-agent |
Behavioral contract for coding agents (always-on via SessionStart hook, and preloaded into the executor, github-actions and gitlab-ci subagents); write-less-code minimalism skill (always-on via hooks); retroactive refactoring (shrink-diff, refactor-repo); usage-limit guard + handoff (usage-limit-handoff); explaining code to a person until it lands (explain-until-understood); whole-repo orientation — explain-codebase into .agents_workspace/CODEBASE_EXPLAINED.md and the repo-tree-mapper agent |
| Plan Build Review | ceh-plan-build-review |
Plan-driven development loop: plan a fullstack app, implement from the plan, review against it |
| Architecture | ceh-architecture |
Living architecture docs (Mermaid diagrams + Key Decisions) and domain modeling (stack-agnostic design) |
| Python Service | ceh-python-service |
FastAPI, asyncpg, PostgreSQL, Alembic, uv, testing, observability, security |
| Python Library | ceh-python-library |
Packaging, public API, semver, uv, testing (no web deps) |
| Web Frontend | ceh-web-frontend |
SvelteKit + React (Vite), Bun, TS style, ESLint/Prettier, Vitest, Playwright, accessibility |
| Scaffolding | ceh-scaffolding |
Per-project-type setup: directory layout + config + .gitignore |
| Git Workflow | ceh-git-workflow |
Commits, branching, PRs, merging, changelog entries, releases, code review, dependency management, plus the merge-flow and release-flow orchestrations |
| Ops | ceh-ops |
Incident response, rollback, deploy pipeline; CI agents |
| Summarize Chat | ceh-summarize-chat |
Structured session summary for LLM handoff |
| Lessons Learned | ceh-lessons-learned |
Session retrospectives into LESSONS_LEARNED.md |
| Blog | ceh-blog |
Interview-driven blog post writing — from rough idea to publication-ready draft |
| Documentation | ceh-documentation |
End-user/operator docs — user guides, runbooks, install/config, troubleshooting; changelog & README maintenance |
| SEO | ceh-seo |
SEO/GEO discoverability for anything internet-exposed — public web pages (meta, structured data, sitemap, llms.txt, rendering) and public-facing text (README first screen, package descriptions, repo topics) |
| Orchestration (experimental) | ceh-orchestration |
Thin-orchestrator mode for cost-optimized multi-step work: plan/delegate-only main session + executor/verifier subagents (and the built-in Explore agent) |
| Business Plan | ceh-business-plan |
Turn a product idea or app plan into a validated business plan via a product-market-fit interview loop — draft, interrogate the weakest assumption, revise until a PMF gate passes |
| Evaluation | ceh-evaluation |
Evaluate a skill or plugin you just wrote — derive its own criteria, measure structure/triggering/content/behavioral lift with evidence, loop fix→re-run until a readiness gate passes |
| Fabled (experimental) | ceh-fabled |
Frontier-grade reasoning discipline for any non-trivial task — deliberate thinking, alternative generation, adversarial self-review, verification, calibrated conviction |
| Advisor (experimental) | ceh-advisor |
Stronger-model second-opinion subagent for decision points, failure loops, irreversible actions, and pre-completion gates — plus hook backstops (destructive-command guard, failure watch) |
| Testing | ceh-testing |
Stack-agnostic testing technique — reproduce-first bug fixes and bisection, systematic test-case design (partitions, boundaries, properties, metamorphic, fuzzing), suite audits (assertions, mutation, flakiness), behavior-preservation checks for refactors, and a pre-completion risk gate |
| Usability Audit | ceh-usability-audit |
Measure whether a non-expert can actually use what you built — cold persona-constrained walkthroughs (novice-walker), a five-question interface audit across web UI/CLI/library/app surfaces, error-message rewrites, and a plain-language pass |
Plugins split on a single axis — use case — so you load exactly what your work needs. They fall into four tiers:
| Tier | Loaded | Plugins |
|---|---|---|
| Scenario bundle | one per situation | ceh-scenario-* — manifest only, no skills; names the set below |
| Cross-cutting | most sessions | ceh-coding-agent, ceh-git-workflow, ceh-testing, plus ceh-fabled and ceh-advisor (experimental) |
| Use-case workflow | per activity | ceh-plan-build-review, ceh-blog, ceh-business-plan, ceh-evaluation, ceh-usability-audit, ceh-documentation, ceh-seo, ceh-ops, ceh-summarize-chat, ceh-lessons-learned, ceh-scaffolding, ceh-orchestration (experimental) |
| Stack / build | per project type | ceh-python-service, ceh-python-library, ceh-web-frontend, ceh-architecture |
Each plugin is self-contained: a foundational standard needed by more than one plugin is duplicated into each rather than extracted into a shared base, so one plugin per use case is all you load. Cross-cutting plugins are the orthogonal tier — they hold a discipline that applies whatever you are building, so they load alongside a use-case plugin rather than instead of one.
| Plugin | Skill | Invoke as | When to use |
|---|---|---|---|
ceh-coding-agent |
Agent Coding Contract | /ceh-coding-agent:agent-coding-contract |
Start of any coding session — core rules, five-step workflow, stop conditions, non-goals |
ceh-coding-agent |
Write Less Code | /ceh-coding-agent:write-less-code |
Every coding session (auto — session-start load + per-turn reinforcement) — the minimalism ladder (YAGNI → stdlib → native → installed dep → one line) |
ceh-coding-agent |
Shrink Diff | /ceh-coding-agent:shrink-diff |
Branch functionally done, before the PR — retroactively apply write-less-code to the accumulated diff vs main |
ceh-coding-agent |
Refactor Repo | /ceh-coding-agent:refactor-repo |
Manual only — propose-then-apply refactor campaign over the whole repo or a named module |
ceh-coding-agent |
Usage Limit Handoff | /ceh-coding-agent:usage-limit-handoff |
Auto via PostToolUse guard hook when 5h or weekly usage crosses the threshold (default 90%) — stop cleanly, write a handoff artifact for the next session, end the turn |
ceh-coding-agent |
Explain Until Understood | /ceh-coding-agent:explain-until-understood |
Explaining a subsystem, design, or diff to the person in the session, starting from the assumption they are new to the project and know nothing about it: a stated floor, foundations first, plain language, verified claims, plain-ASCII pictures for structure and time, a numbered walk of one real case, and the escalation ladder when an explanation misses |
ceh-plan-build-review |
Plan Fullstack App Iteratively | /ceh-plan-build-review:plan-fullstack-app-iteratively |
Planning one release at a time — a greenfield skeleton or the next iteration |
ceh-plan-build-review |
Plan Fullstack App to MVP | /ceh-plan-build-review:plan-fullstack-app-to-mvp |
Planning the complete build to a working MVP in one session |
ceh-plan-build-review |
Implement From Plan | /ceh-plan-build-review:implement-from-plan |
Implementing a SKELETON.md or ITER_NN.md planning document |
ceh-plan-build-review |
Review Against Plan | /ceh-plan-build-review:review-against-plan |
Auditing implementation against a SKELETON.md or ITER_NN.md planning document |
ceh-plan-build-review |
Patch Built Version | /ceh-plan-build-review:patch-built-version |
Patching an already-implemented version — a small, non-feature change recorded as a patch ITER_NN.md; routes features to the iterative planner |
ceh-architecture |
Document Architecture | /ceh-architecture:document-architecture |
Writing/updating the living ARCHITECTURE.md — a 3-second Overview, Mermaid diagrams, a Key Decisions log |
ceh-architecture |
Domain Modeling | /ceh-architecture:domain-modeling |
Designing entities, IDs, status fields, or layer boundaries |
ceh-python-service |
FastAPI | /ceh-python-service:fastapi |
Writing route handlers, DI, middleware, exception hierarchy, or REST API design |
ceh-python-service |
asyncpg | /ceh-python-service:asyncpg |
Writing database queries, transactions, tenant isolation, or connection pool config |
ceh-python-service |
PostgreSQL | /ceh-python-service:postgresql |
Designing a schema, choosing column types, or adding indexes |
ceh-python-service |
Alembic | /ceh-python-service:alembic |
Creating or running database migrations; migration deploy safety |
ceh-python-service |
Python Service Environment | /ceh-python-service:python-service-environment |
Setting up uv/pyproject.toml, writing type hints, configuring ruff/mypy |
ceh-python-service |
Python Service Testing | /ceh-python-service:python-service-testing |
Writing Python unit or integration tests |
ceh-python-service |
Python Observability | /ceh-python-service:python-observability |
Adding structlog logging, metrics, health checks, or correlation IDs |
ceh-python-service |
Python Security | /ceh-python-service:python-security |
Secrets management, CORS, rate limiting, or input validation |
ceh-python-library |
Packaging | /ceh-python-library:packaging |
Build backend, src layout, wheels/sdist, publishing to PyPI |
ceh-python-library |
Public API | /ceh-python-library:public-api |
Defining __all__, changing a public signature, classifying a semver bump |
ceh-python-library |
Python Library Environment | /ceh-python-library:python-library-environment |
Setting up uv/pyproject.toml for a library (no web deps) |
ceh-python-library |
Python Library Testing | /ceh-python-library:python-library-testing |
Writing unit and public-API tests for a library |
ceh-web-frontend |
Environment | /ceh-web-frontend:environment |
Bun/Vite setup, TypeScript style, ESLint/Prettier, type config |
ceh-web-frontend |
SvelteKit | /ceh-web-frontend:sveltekit |
Working on Svelte routes, stores, components, or the API client |
ceh-web-frontend |
React + Vite | /ceh-web-frontend:react-vite |
Working on React components, hooks, routing, or Vite config |
ceh-web-frontend |
Frontend Testing | /ceh-web-frontend:frontend-testing |
Writing Vitest, Testing Library, MSW, or Playwright tests |
ceh-web-frontend |
Accessibility | /ceh-web-frontend:accessibility |
Writing component markup (Svelte or React) |
ceh-web-frontend |
UI Design | /ceh-web-frontend:ui-design |
Frontend UI visual design — layout, hierarchy, navigation, states, finishing recipes for first-pass polish, plus theme/brand from bundled templates |
ceh-scaffolding |
Scaffold Python Service | /ceh-scaffolding:scaffold-python-service |
Starting a FastAPI/Python web service repo |
ceh-scaffolding |
Scaffold Python Library | /ceh-scaffolding:scaffold-python-library |
Starting a distributable Python library/package |
ceh-scaffolding |
Scaffold Web Frontend | /ceh-scaffolding:scaffold-web-frontend |
Starting a SvelteKit or React + Vite frontend |
ceh-scaffolding |
Scaffold Fullstack Web | /ceh-scaffolding:scaffold-fullstack-web |
Starting a fullstack web app (service + frontend in one repo) |
ceh-git-workflow |
Branch | /ceh-git-workflow:branch |
Creating or naming a branch |
ceh-git-workflow |
Commit | /ceh-git-workflow:commit |
Writing a commit message or staging changes |
ceh-git-workflow |
Open PR | /ceh-git-workflow:open-pr |
Opening a pull request, writing a PR description, checking the definition of done, or enabling auto-merge on repos that allow it |
ceh-git-workflow |
Merge | /ceh-git-workflow:merge |
Merging a PR (immediate or auto-merge) or a local branch into main, then deleting the branch afterward |
ceh-git-workflow |
Hotfix | /ceh-git-workflow:hotfix |
Executing a critical production fix |
ceh-git-workflow |
Release | /ceh-git-workflow:release |
Tagging a release or bumping a version |
ceh-git-workflow |
Code Review | /ceh-git-workflow:code-review |
Reviewing a PR or leaving review comments |
ceh-git-workflow |
Dependency Management | /ceh-git-workflow:dependency-management |
Adding or upgrading a package |
ceh-git-workflow |
Update Changelog | /ceh-git-workflow:update-changelog |
Generate or update CHANGELOG.md, write release notes, or log a change under [Unreleased] |
ceh-git-workflow |
Merge Flow | /ceh-git-workflow:merge-flow |
Land the branch you are on in one pass — changelog under [Unreleased] → README → commit → PR → merge → cleanup, with no version bump and no tag |
ceh-git-workflow |
Release Flow | /ceh-git-workflow:release-flow |
Ship a complete release in one pass — version bump → changelog → README → CLAUDE.md → PR → merge → tag → release, sequencing the skill that owns each step |
ceh-ops |
Deploy | /ceh-ops:deploy |
Building/promoting images, staging→prod, post-deploy health checks, change classification |
ceh-ops |
Incidents | /ceh-ops:incidents |
Responding to a production incident or writing a post-mortem |
ceh-ops |
Rollback | /ceh-ops:rollback |
Deciding to roll back a deployment or recovering from a failed migration |
ceh-summarize-chat |
Summarize Chat | /ceh-summarize-chat:summarize-chat |
Summarizing the current session for handoff to a future LLM session |
ceh-lessons-learned |
Lessons Learned | /ceh-lessons-learned:lessons-learned |
Extracting lessons learned from the current session into LESSONS_LEARNED.md |
ceh-blog |
Blog Interviewer | /ceh-blog:blog-interviewer |
Turn a rough idea, project, or experience into a compelling, publishable blog post |
ceh-blog |
Blog Writer | /ceh-blog:blog-writer |
Draft straight from existing notes, bullets, or outline — no interview |
ceh-blog |
Blog Editor | /ceh-blog:blog-editor |
Diagnose and polish an existing draft — diagnosis first, then a full revised version |
ceh-blog |
Blog Repurpose | /ceh-blog:blog-repurpose |
Adapt a finished post into Twitter/X thread, LinkedIn post, TL;DR, or newsletter blurb |
ceh-documentation |
User & Operator Guide | /ceh-documentation:user-operator-guide |
Writing a user guide, operator runbook, getting-started/install/config guide, or troubleshooting reference |
ceh-documentation |
Update README | /ceh-documentation:update-readme |
Refresh README after a significant change (new feature, changed install steps, new API surface) |
ceh-coding-agent |
Explain Codebase | /ceh-coding-agent:explain-codebase |
Go through a whole repo and write what each component does, how they connect, and key flows into git-ignored .agents_workspace/CODEBASE_EXPLAINED.md (per-file detail only on request) |
ceh-seo |
Web Discoverability | /ceh-seo:web-discoverability |
Shipping a public web page/route — head checklist, sitemap/robots/llms.txt, JSON-LD, SSR/prerender, GEO citation rules |
ceh-seo |
Text Discoverability | /ceh-seo:text-discoverability |
Writing public-facing repo/package text — README first screen, one-liner, GitHub topics, PyPI/npm descriptions and keywords |
ceh-orchestration |
Orchestrate | /ceh-orchestration:orchestrate |
Decompose and delegate a big multi-step task — plan/delegate-only main session, cheap isolated workers, to cap context/token cost |
ceh-business-plan |
Develop Business Plan | /ceh-business-plan:develop-business-plan |
Draft a business plan proactively from app plans or a product idea, then loop interview→revise until the product-market-fit readiness gate passes |
ceh-evaluation |
Evaluate Skill | /ceh-evaluation:evaluate-skill |
Evaluate a skill or plugin you wrote — derive its criteria, measure structure/triggering/content/behavioral lift with evidence, loop fix→re-run until a 6-point gate passes |
ceh-evaluation |
Evaluate Skill — Lite | /ceh-evaluation:evaluate-skill-lite |
Fast dev-loop check — structure + triggering (single pass) + content only; skips behavioral lift, reports a partial 4/6 gate for cheap iteration before the full ship verdict |
ceh-fabled |
Fabled | /ceh-fabled:fabled |
Any non-trivial task with more than one plausible answer — deliberate reasoning, alternatives, adversarial self-review, verification, and calibrated conviction |
ceh-fabled |
Fabled Plan Review | /ceh-fabled:fabled-plan-review |
Review an existing plan against frontier-grade planning discipline — problem fidelity, alternatives, decomposition, pre-mortem, verifiability — verdict plus concrete fixes |
ceh-fabled |
Fabled Stuck | /ceh-fabled:fabled-stuck |
Escape a failure loop after repeated failed fixes — freeze, inventory attempts, attack their shared assumption, re-derive the diagnosis from evidence, probe before fixing |
ceh-testing |
Test a Bug Fix | /ceh-testing:test-a-bug-fix |
Fixing a bug, crash, regression, or incident — reproduce-first: failing test before the fix, prove the test is coupled to it, bisect on it when the behavior used to work |
ceh-testing |
Design Test Cases | /ceh-testing:design-test-cases |
Choosing which inputs and scenarios to cover — partitions, boundaries, decision tables, state transitions, pairwise, properties, metamorphic relations, fuzzing, forced dependency failure |
ceh-testing |
Audit Test Suite | /ceh-testing:audit-test-suite |
Deciding whether a passing suite would actually catch a defect — assertion audit, delete-the-code check, mutation testing on the diff, flakiness, branch coverage |
ceh-testing |
Verify Behavior Preserved | /ceh-testing:verify-behavior-preserved |
Before a no-behavior-change edit (refactor, extraction, dependency/runtime upgrade, port) — characterization tests, golden files, differential run |
ceh-testing |
Close Test Risk Gaps | /ceh-testing:close-test-risk-gaps |
Pre-completion gate — triage concurrency/idempotency, contract drift, performance regression, authorization, and migration/rolling-deploy compatibility; skip each class explicitly when its trigger does not fire |
ceh-usability-audit |
First-Run Walkthrough | /ceh-usability-audit:first-run-walkthrough |
Can a stranger reach first success — install, sign-up, setup, onboarding; cold persona walkers, ranked by observed stalls, milestones capped by an action budget set before the walk, looped to a 5-point gate |
ceh-usability-audit |
Audit Interface | /ceh-usability-audit:audit-interface |
They are already in — the five questions every web UI/CLI/API/screen must answer unasked, a reject-on-sight anti-pattern sweep, the naming test, and the persona battery |
ceh-usability-audit |
Audit Error Messages | /ceh-usability-audit:audit-error-messages |
Anything a user reads when something goes wrong — the three-part rule (what happened, what was wrong, what to do next) over every user-reachable string |
ceh-usability-audit |
Plain Language Pass | /ceh-usability-audit:plain-language-pass |
Labels, help text, empty states, confirmation dialogs, onboarding copy — vocabulary floor, sentence rules, and an explicit never-simplify list |
ceh-fabled |
Fabled Voice | /ceh-fabled:fabled-voice |
Always-on via SessionStart hook — deliver in fable's writing style — finding-first progress lines, verdict-first advisory answers closing on a calibration, and reports built from bold inline labels, hard numbers, a validated/not-validated ledger, and a standing offer |
Agents run autonomously for a defined task and hand results back to the parent session.
Plugin-agent limitation: every agent here ships inside a plugin. Claude Code ignores the
permissionMode,hooks, andmcpServersfrontmatter fields on plugin subagents (for security reasons), so no agent in this repo sets them. These agents inherit the permission context of your session and prompt for edit/write permissions accordingly. To avoid the prompts, put the session inacceptEdits(Shift+Tab) before dispatching — a parentacceptEditsorbypassPermissionstakes precedence and is inherited — or addpermissions.allowrules insettings.json. See the subagents docs.Background tool filter: subagents run in the background by default, and a background subagent keeps only
Read,Grep,Glob,Bash,PowerShell,Edit,Write,NotebookEdit,WebFetch,WebSearch,TodoWrite,Skill,ToolSearch,EnterWorktree,ExitWorktree,Monitor,TaskStop,SendMessage, andArtifact. Anything else is stripped silently, even when named intools:.AskUserQuestionis removed from every subagent, so no agent here can stop to ask you a question.
| Plugin | Agent | Invoke as | When to use |
|---|---|---|---|
ceh-coding-agent |
Repo Tree Mapper | /ceh-coding-agent:repo-tree-mapper |
Map or document a repository's structure; auto-triggers on orientation requests |
ceh-python-service |
Python Unit Tester | /ceh-python-service:python-unit-tester |
Write isolated pytest unit tests for functions, classes, or modules with mocked dependencies |
ceh-python-service |
Python Integration Tester | /ceh-python-service:python-integration-tester |
Write tests for real component interactions — DB, HTTP between internal services, service boundaries |
ceh-python-service |
Python System Tester | /ceh-python-service:python-system-tester |
Write full end-to-end / acceptance tests that exercise the entire application stack |
ceh-web-frontend |
TS Unit Tester | /ceh-web-frontend:ts-unit-tester |
Write isolated Vitest unit tests for functions, classes, and modules with mocked dependencies |
ceh-web-frontend |
TS Integration Tester | /ceh-web-frontend:ts-integration-tester |
Write tests wiring real stores, MSW handlers, and multiple components together |
ceh-web-frontend |
TS System Tester | /ceh-web-frontend:ts-system-tester |
Write Playwright E2E tests that exercise the full running stack as a real user would |
ceh-testing |
Test Suite Auditor | /ceh-testing:test-suite-auditor |
Run the slow, high-output suite audit out of session — mutation testing on the diff, flakiness and isolation runs — and hand back a ranked read-only report (Sonnet) |
ceh-git-workflow |
Commit Author | /ceh-git-workflow:commit-author |
Create one commit in an isolated subagent; derives the change from git diff, pass only the why (Sonnet, medium effort) |
ceh-git-workflow |
PR Opener | /ceh-git-workflow:pr-opener |
Push the branch and open the PR in an isolated subagent; queues auto-merge where allowed (Sonnet, medium effort) |
ceh-git-workflow |
Branch Merger | /ceh-git-workflow:branch-merger |
Merge a PR or local branch into main and clean up, gate-checked, in an isolated subagent (Sonnet, medium effort) |
ceh-git-workflow |
Release Cutter | /ceh-git-workflow:release-cutter |
Tag main and publish the release in an isolated subagent; bump commit only if not landed (Sonnet, medium effort) |
ceh-ops |
GitHub Actions | /ceh-ops:github-actions |
Create or fix GitHub Actions workflows, jobs, matrix builds, OIDC, reusable workflows |
ceh-ops |
GitLab CI | /ceh-ops:gitlab-ci |
Create or fix .gitlab-ci.yml pipelines, DAG stages, rules, protected variables, runners |
ceh-orchestration |
Executor | /ceh-orchestration:executor |
Implement a single scoped task: code changes, edits, multi-step work (Sonnet) |
ceh-orchestration |
Verifier | /ceh-orchestration:verifier |
Check an executor's output against acceptance criteria — PASS/FAIL only (Haiku) |
ceh-usability-audit |
Novice Walker | /ceh-usability-audit:novice-walker |
Walk a target cold under one persona toward one goal and report where it stalled — may not use what it already knows about how such tools usually work (Sonnet, read-only) |
ceh-advisor |
Advisor | /ceh-advisor:ceh-advisor |
Verdict-first second opinion before an architectural commit, after 2+ failed fixes, before an irreversible action, or before declaring a complex task done (Opus, high effort); requires a handoff block — also hard-triggered by the plugin's hooks on destructive commands and failure streaks |
python3 on PATH — required by the hooks in ceh-advisor and ceh-coding-agent
(ceh-fabled's SessionStart hook needs only bash)
(stdlib only, no packages). Every other plugin works without it.
This matters most for ceh-advisor: its destructive-command guard fails closed, so without
python3 it blocks every rm -rf, git push --force, terraform destroy, and similar until the
interpreter is available. Install with winget install Python.Python.3.12 / brew install python
/ apt install python3.
/plugin marketplace add cheneeheng/agent-skills
Prefer a scenario bundle (see Start here) — it pulls its members in automatically. Install individual plugins only when you want a set no bundle describes:
/plugin install ceh-coding-agent@ceh-plugins --scope user
/plugin install ceh-plan-build-review@ceh-plugins --scope user
/plugin install ceh-git-workflow@ceh-plugins --scope user
/plugin install ceh-architecture@ceh-plugins --scope user
/plugin install ceh-python-service@ceh-plugins --scope user
/plugin install ceh-python-library@ceh-plugins --scope user
/plugin install ceh-web-frontend@ceh-plugins --scope user
/plugin install ceh-scaffolding@ceh-plugins --scope user
/plugin install ceh-ops@ceh-plugins --scope user
/plugin install ceh-summarize-chat@ceh-plugins --scope user
/plugin install ceh-lessons-learned@ceh-plugins --scope user
/plugin install ceh-blog@ceh-plugins --scope user
/plugin install ceh-documentation@ceh-plugins --scope user
/plugin install ceh-orchestration@ceh-plugins --scope user
/plugin install ceh-business-plan@ceh-plugins --scope user
/plugin install ceh-evaluation@ceh-plugins --scope user
/plugin install ceh-fabled@ceh-plugins --scope user
/plugin install ceh-advisor@ceh-plugins --scope user
/plugin install ceh-testing@ceh-plugins --scope user
/plugin install ceh-usability-audit@ceh-plugins --scope user
Some of those are already dependencies of others: installing ceh-ops brings ceh-coding-agent,
and each stack plugin brings ceh-testing.
Or install all at once using --scope project for project-specific installs.
/help
The ceh-*: skills should appear in the skills list.
Clone this repo and point Claude Code at the plugin subdirectories directly:
git clone https://github.com/cheneeheng/agent-skills.git ~/agent-skillsThen add plugin paths to your Claude Code settings (~/.claude/settings.json):
{
"plugins": [
{ "path": "~/agent-skills/plugins/ceh-coding-agent" },
{ "path": "~/agent-skills/plugins/ceh-plan-build-review" },
{ "path": "~/agent-skills/plugins/ceh-git-workflow" },
{ "path": "~/agent-skills/plugins/ceh-architecture" },
{ "path": "~/agent-skills/plugins/ceh-python-service" },
{ "path": "~/agent-skills/plugins/ceh-python-library" },
{ "path": "~/agent-skills/plugins/ceh-web-frontend" },
{ "path": "~/agent-skills/plugins/ceh-scaffolding" },
{ "path": "~/agent-skills/plugins/ceh-ops" },
{ "path": "~/agent-skills/plugins/ceh-summarize-chat" },
{ "path": "~/agent-skills/plugins/ceh-lessons-learned" },
{ "path": "~/agent-skills/plugins/ceh-blog" },
{ "path": "~/agent-skills/plugins/ceh-documentation" },
{ "path": "~/agent-skills/plugins/ceh-orchestration" },
{ "path": "~/agent-skills/plugins/ceh-business-plan" },
{ "path": "~/agent-skills/plugins/ceh-evaluation" },
{ "path": "~/agent-skills/plugins/ceh-fabled" },
{ "path": "~/agent-skills/plugins/ceh-advisor" },
{ "path": "~/agent-skills/plugins/ceh-seo" },
{ "path": "~/agent-skills/plugins/ceh-testing" },
{ "path": "~/agent-skills/plugins/ceh-usability-audit" }
]
}| Tool | Path | Purpose |
|---|---|---|
| skills-sync | tools/skills-sync/ |
Copy individual skills (from this repo or any other) into a project's .claude/skills/ directory — install, update, add, remove, list. Python, bash, PowerShell, and browser-based HTML implementations. |
| validate-plugins | tools/validate-plugins/ |
Repo-integrity checker run by CI (.github/workflows/validate.yml): validates plugin manifests, skill/agent frontmatter (including a 1024-char cap on description), file and skill references, and script syntax. Stdlib-only Python. |
tools/ holds standalone meta-tooling that isn't itself a ceh-* plugin, skill, or agent — see
tools/skills-sync/README.md for usage.