Help us improve
Share bugs, ideas, or general feedback.
Share bugs, ideas, or general feedback.
Share bugs, ideas, or general feedback.
By punt-labs
Punt Labs standards enforcement — LLM-powered reconciliation of workflows, configs, and project structure
npx claudepluginhub punt-labs/claude-plugins --plugin punt[DEV] Check compliance using the working tree
Check compliance against Punt Labs standards
[DEV] Execute an automation playbook using the working tree
Execute an automation playbook
[DEV] Convert this plugin's commands to Cursor skills using the working tree
Share bugs, ideas, or general feedback.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Project infrastructure standards - comprehensive project configuration for pre-commit, CI/CD, Docker, testing, linting, formatting, and more
Reorganizes project structure by cleaning root clutter, creating logical folder hierarchies, and moving files to optimal locations. Tracks dependencies and fixes broken imports/paths. Use PROACTIVELY when project structure becomes unwieldy or needs architectural cleanup.
AI-First SDLC — zero-debt development with validators, enforcement, and workflows
AI-powered development workflow automation - Phase-based planning, implementation orchestration, preflight code quality checks with security scanning, ship-it workflow, and development principles generator for CLAUDE.md
Context-Driven Development plugin that transforms Claude Code into a project management tool with structured workflow: Context → Spec & Plan → Implement
Validation and quality enforcement for GitLab CI/CD pipeline configurations.
Amazon Working Backwards PR/FAQ process — generate professional LaTeX documents for product discovery and decision-making
Create, validate, and test formal Z specifications for stateful systems using fuzz and probcli
Autonomous agent daemon with cryptographic owner control. Email communication via IMAP/SMTP with PGP trust model.
Identity binding for humans and AI agents — voice, email, GitHub, writing style, personality.
UNIX-style team communication for Claude Code: /who, /finger, /plan, /write, /read, /mesg, /tty, /last, /wall, /talk
Standards, compliance tooling, and Claude Code plugin development patterns.
Punt-kit serves two purposes. For Punt Labs projects, it is the standards authority — coding conventions, CI templates, naming rules, and an audit checklist that the punt CLI enforces automatically. For Claude Code plugin developers, it is a pattern library — ten reusable design patterns extracted from shipping plugins, plus nine standards documents covering Python, CLI, shell, distribution, and more.
Individual projects reference these standards from their CLAUDE.md rather than duplicating them. The standards/ directory is the source of truth; patterns/ documents the design patterns; AGENTS.md has the full audit checklist.
Platforms: macOS, Linux
curl -fsSL https://raw.githubusercontent.com/punt-labs/punt-kit/e36a3bf/install.sh | sh
uv tool install punt-kit
| Command | What it does |
|---|---|
punt init | Scaffold a new Punt Labs project (CI, linter config, CLAUDE.md, permissions, beads) |
punt audit | Check compliance against standards (read-only, no file changes) |
punt auto <target> | Render and merge managed sections in project files (claude, makefile, settings). Supports --dry-run |
punt release [version] | 11-phase release workflow (preflight → PyPI → cross-repo propagation). Supports --resume-from and --dry-run |
punt pii | Scan repo for PII (emails, home paths, hostnames). Supports --staged for pre-commit |
punt doctor | Check installation health (Python, uv, ruff, mypy, pyright) |
punt status | Show detected project type, standards version, and beads state |
punt version | Print version (punt 0.8.0) |
All commands accept these flags:
| Flag | What it does |
|---|---|
--json | Output as JSON (machine-readable) |
--verbose / -v | Verbose output |
--quiet / -q | Suppress non-essential output (errors only) |
# In any Punt Labs project directory:
punt init # Generate missing files
punt audit # Check compliance
punt doctor # Check installation health
punt status # Show project summary
punt pii # Scan for PII leaks
punt pii --staged # Pre-commit: scan only staged files
punt --json status # JSON output for scripting
punt init /path # Target a specific directory
punt init detects the project type (Python, Node.js, Swift, docs) and scaffolds everything a Punt Labs project needs to pass punt audit:
[tool.ruff], [tool.mypy], [tool.pyright], [tool.pytest] sections merged into pyproject.toml.claude/settings.json with tool allowlists and standard deny rules.claude/ exceptions appended to .gitignorebd initAfter scaffolding, it reports manual steps that require GitHub access (branch protection rulesets, Copilot code review, Dependabot). Init is idempotent — re-running it updates existing files without overwriting customizations.
punt audit checks all standards without modifying files. It uses gh api to check GitHub-side settings when authenticated.
punt pii scans tracked files for personally identifiable information — email addresses, home directory paths (/Users/X/, /home/X/), and .local hostnames. Configure allow/deny lists in [tool.punt.pii] in pyproject.toml.
Punt-kit is also a Claude Code plugin (punt@punt-labs on the marketplace). The plugin surfaces CLI commands as slash commands and adds LLM-driven capabilities that are inherently prompt-driven.