By zeveck
Agent-discipline skill framework: plans, fix-issues, draft-plan, run-plan, land-pr, and more.
Full pipeline: decompose a broad goal into sub-plans, draft each with adversarial review, then execute all of them autonomously via /run-plan. One command, walk away.
Decompose a broad goal into a sequence of executable sub-plans. Researches the domain, identifies sub-problems and dependencies, produces a meta-plan whose phases each delegate to /run-plan.
Audit what THIS session said it would do vs. what's actually shipped. Verifies session-mentioned items against ground truth (git, PRs, plans, worktrees), not conversation memory — items may have been completed in another session.
Generate a project briefing: worktree status, open checkboxes, recent commits. Modes: summary (default), report, verify, current, worktrees. Period: 1h, 6h, 24h, 2d, 7d.
Post-PR-merge local normalization: fetch-and-prune origin, switch off merged feature branches, pull main, and delete local branches whose upstream is gone, whose PR has merged, or whose tip is fully contained in main (0 commits ahead). Preview by default — run `/cleanup-merged apply` to execute. `local` (default) / `remote` / `all` pick the scope; pass explicit branch names to narrow the candidate set. `--force` overrides the merged-check + unpushed guard for branches you explicitly name. Protected branches from config are NEVER deleted (even with `--force`) — they are always skipped.
Matches all tools
Hooks run on every tool call, not just specific ones
Executes bash commands
Hook triggers when Bash tool is used
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.
Modifies files
Hook triggers on file write and edit operations
Modifies files
Hook triggers on file write and edit operations
23 core skills that plan, build, test, fix, and ship — so one developer can run a full engineering team. (21 user-facing slash commands, 2 internal helpers, and a battery of safety hooks.)
Z Skills encodes hard-won lessons from real agent failures into reusable
prompt files. Each skill is a .claude/skills/<name>/SKILL.md file that
teaches Claude Code how to drive a specific workflow with the discipline
that prevents the most common AI agent failure modes: skipping
verification, weakening tests, deferring hard parts, and shipping broken
code.
Read the docs in the in-browser docs viewer or here in the repo.
New users may want to explore Installing zskills, the Workflows guide, or the Skill reference.
For a higher level view see the ZSkills presentation →.



By default zskills installs as a Claude Code plugin. From inside a Claude Code session in your project:
/plugin marketplace add zeveck/zskills
/plugin install zs@zskills
Then restart Claude Code (or /clear) so we can finish setup — the
SessionStart hook materialises zskills' consumer-side artifacts on the next
session start (plugins cannot write at install time). Finally, run:
/update-zskills
to confirm the install and check your environment (it reports whether git, Python, and gh are present). Until you run it, a one-line greeting on startup reminds you.
zskills needs git and bash. Its worktree, commit, and cherry-pick machinery is built on
git, and its hooks and helper scripts run underbash./update-zskillsreports whether git (and gh, for PR-mode landing) are present; bash is assumed by every shipped script.
This is not a pip/npm package — do not pip install or npm install it.
The repo contains prompt files and scripts.
There's a more flexible direct install alternative — instead of the plugin,
it copies the skill source straight into your repo's .claude/. Same skills,
same behavior; the visible differences are the slash prefix and where the source
lives. Pick one — running both at once isn't a supported end-state (run
scripts/switch-install-path.sh to consolidate).
| Plugin | /update-zskills (direct) | |
|---|---|---|
| Install | /plugin marketplace add zeveck/zskills/plugin install zs@zskills | clone + copy skills, then /update-zskills install |
| Update | /plugin marketplace update | /update-zskills install |
| Slash prefix | /zs:run-plan, /zs:do | bare /run-plan, /do |
| Skill source in your repo | no (plugin-managed) | yes (copied into .claude/) |
For more details see Installing zskills.
/update-zskills installTell your agent (copy-paste):
Install zskills from github.com/zeveck/zskills — see repo for directions
A capable agent will clone the repo, copy the skills into your
project, and run /update-zskills to complete setup. Full manual steps
are below if you prefer to drive it yourself.
Clone the repo (if not already cloned):
git clone https://github.com/zeveck/zskills.git /tmp/zskills
Copy skills from the clone into your project:
mkdir -p .claude/skills
cp -r /tmp/zskills/skills/* .claude/skills/
Run /update-zskills to complete setup. This is the important
step — it creates CLAUDE.md with auto-detected project settings,
installs hooks and scripts, registers hooks in settings.json,
writes .claude/zskills-config.json, verifies dependencies, and
reports any gaps. On a greenfield project it will ask you a single
question: which landing mode you want (see below).
That's it. /update-zskills handles everything beyond the initial skill
copy.
On a fresh project /update-zskills will prompt:
How should /run-plan land changes?
(1) cherry-pick — each phase squash-lands directly to main (simple, solo)
(2) locked-main-pr — plans become feature branches + PRs, CI, auto-merge
(locked main, shared repo)
(3) direct — work on main, no worktree isolation (minimal, risky)
You can also pass the preset directly — no prompt — when you already know what you want:
/update-zskills cherry-pick
/update-zskills locked-main-pr
/update-zskills direct
npx claudepluginhub zeveck/zskills --plugin zsAgent-discipline skill framework: plans, fix-issues, draft-plan, run-plan, land-pr, and more.
Permanent coding companion for Claude Code — survives any update. MCP-based terminal pet with ASCII art, stats, reactions, and personality.
Intelligent prompt optimization: injects the right context at the right moment so Claude lands a better first output. Clarifies vague prompts with research-based questions, plus targeted nudges for approach selection, plan readability, workflow routing, background execution, subagent routing, output readability, user-decision questions, and plan-mode assessment
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Memory compression system for Claude Code - persist context across sessions