pm-manager

by wei63wVerified

Know what to fix next — local .pm governance skill pack for AI coding agents (Spec Kit–inspired).

85
Stars
0
Forks
Python
Language
8/24/2026
Added
View on GitHubDownload ZIP

⚠️ Third-Party Software Notice

This skill is third-party open-source software developed and hosted independently on GitHub. SkillTip is an informational directory and does not control or maintain the underlying repository. Any security checks displayed are automated and limited in scope. Review the source code before installing.

Read the Terms of Service

Installation

Add to your Claude Code skills directory:

# Add to your Claude Code skills
git clone https://github.com/wei63w/pm-manager

Getting Started

Guides for using skills like pm-manager.

Security Report

Verified

Last scanned: —

{
  "status": "PASSED",
  "issues": []
}

README.md

📋 PM Manager
Know what to fix next — with any AI coding agent.

Release Version Downloads Commit activity License skills.sh

An open source skill pack for local project governance — init a .pm/ workbench, keep a daily Top3, triage pasted logs, and run release-ready health scans. Inspired by Spec Kit’s command-template + multi-agent adapter model. Works alongside Spec Kit: Spec Kit drives what to build; PM Manager drives project health and what’s next.


Table of Contents

What is PM Manager?

Most AI coding sessions jump straight into code. PM Manager flips the day-to-day loop: keep a local, auditable .pm/ board for health, todos, incidents, and release gates — then let the agent recommend at most three things to do today.

It is not a cloud PM suite and not a replacement for Jira/Linear. It is a developer-side governance workbench that lives in your repo’s local .pm/ directory (git-excluded), driven by slash commands / skills the same way Spec Kit drives /speckit.*.

Get Started

1. Install the Agent Skill (skills.sh)

npx skills@latest add wei63w/pm-manager

Or copy skills/pm-manager into your tool’s skills directory:

ToolPath
Cursor~/.cursor/skills/pm-manager/
Claude Code~/.claude/skills/pm-manager/
Codex~/.agents/skills/pm-manager/

2. Install the CLI (optional, cross-platform)

Requires uv and Python 3.11+. The CLI scaffolds .pm/ and installs per-agent adapters.

# From GitHub (recommended)
uv tool install pm-manager-cli --from git+https://github.com/wei63w/pm-manager.git

# Or from a local checkout
uv tool install --editable .

Verify:

pm version
# or
pm-manager version

Upgrade later:

uv tool upgrade pm-manager-cli
# or reinstall from git
uv tool install pm-manager-cli --force --from git+https://github.com/wei63w/pm-manager.git

3. Bootstrap a project

In your application repository (not required to keep this pack checked out):

cd /path/to/your-app

# Create .pm/ + install Cursor & Claude adapters (default)
# Also runs non-interactive: npx skills add wei63w/pm-manager -y
# (skills.sh indexing; requires Node.js/npx — skipped with a warning if missing)
pm init

# Cursor only
pm init --agent cursor

# Claude Code only
pm init --agent claude

# Scaffold .pm/ only (no agent files, no skills.sh)
pm init --scaffold-only

# Skip skills.sh / npx step
pm init --no-skills-sh

Refresh adapters without re-scaffolding:

pm install --agent all
pm install --no-skills-sh   # adapters only
pm check

Windows / macOS / Linux all use the same commands. Legacy PowerShell helpers under scripts/powershell/ still work, but pm init is preferred.

4. Initialize governance in your agent

Open Cursor / Claude Code in the project and run:

/pm-init
  • Detects new vs existing projects
  • Discovers Spec Kit .specify/memory/constitution.md (and other charter candidates) when present
  • For empty projects, accepts a one-line intent to generate an outline
  • Complements the filesystem scaffold from pm init (agent fills charter/overview)

5. Check today's Top3

/pm-status

You’ll get a short health line and up to three actionable todos. Claim one with /pm-next, close it with /pm-done TODO-001.

6. Triage an incident

Paste a stack trace or log into the chat:

/pm-fix

Conversation paste is a first-class evidence source (no need to save a file first).

7. Pre-release full scan

/pm-all

Runs gated full governance with a noise-filtered summary (blocking + high by default). Use --verbose for the long list.

After the scan, open .pm/dashboard/index.html in a browser for the visual dashboard (KPI, charts, module risk). Or read overview.md in the IDE. Rebuild anytime with pm dashboard.

CLI Reference

CommandDescription
pm versionPrint CLI version
pm init [path]Create .pm/ + install agent adapters (+ skills.sh)
`pm init --agent cursorclaude
pm init --scaffold-onlyOnly create .pm/
pm init --no-skills-shSkip npx skills add wei63w/pm-manager
pm install --agent …Refresh adapters without scaffolding
pm install --no-skills-shRefresh adapters only
pm check [path]Show whether .pm/ and adapters exist
pm dashboard [path]Rebuild .pm/dashboard/ from module findings/todos
pm arch [path]Scan project → Mermaid diagrams under .pm/architecture/

pm-manager is an alias of pm.

Supported AI Coding Agents

AgentInstall pathHow you invoke
Cursor.cursor/skills/pm-manager (+ optional per-command skills under adapters/cursor/skills)/pm-init, /pm-status, … or natural language ("what should I do today")
Claude Code.claude/commands/pm-*.md/pm-init, /pm-status, …
Other skill hostsUse skills/pm-manager/SKILL.md as a router into templates/commands/Follow host skill conventions

Natural-language routing (when slash commands are unavailable) is documented in skills/pm-manager/memory/ROUTING.md.

Available Slash Commands

Daily commands (remember these)

CommandAgent skillDescription
/pm-initpm-initCreate .pm/, detect lifecycle, discover charter sources
/pm-statuspm-statusHealth summary + Today's Top3 (primary daily entry)
/pm-nextpm-nextClaim the next todo (in_progress)
/pm-donepm-doneClose TODO-xxx, refresh overview
/pm-fixpm-fixParse pasted logs/stacks into bugs + todos
/pm-allpm-allFull gated scan; rebuilds .pm/dashboard/ aggregate
/pm-archpm-archGenerate architecture + flow Mermaid diagrams from repo

Planning & export

CommandAgent skillDescription
/pm-outlinepm-outlineGenerate outline + draft charter from intent
/pm-charterpm-chartercreate / import / discover / approve / skip charter
/pm-exportpm-exportDesensitized markdown summary for share / machine switch

Internal (usually via /pm-all)

CommandAgent skillDescription
/pm-discoverpm-discoverDeep-scan all enabled modules

Command prompts live in skills/pm-manager/templates/commands/ with Spec Kit–style frontmatter and handoffs.

How it relates to Spec Kit

Spec KitPM Manager
.specify/.pm/ (local, not committed)
/speckit.constitution/pm-charter + auto-discover constitution
Spec → plan → tasks → implementStatus → Top3 → done / fix / all
Spec-driven feature deliveryGovernance-driven project health

They can run in the same repo. PM Manager will read Spec Kit artifacts when present; it does not replace Spec-Driven Development.

Core Philosophy

  • Simple daily UX — few entry commands; Top3 over long finding dumps
  • Local & auditable — evidence and todos stay in .pm/ with redaction
  • On-demand scans — no background daemons; paste / --path / registered sources
  • Safe defaults — report first; never auto-write app code/SQL/cloud without confirmation
  • Charter-aware (optional) — without a charter you get technical debt scans; with one you get scoped “unreasonable” checks with confidence

Daily Workflow

MomentCommand
First time in a repo/pm-init
Morning / “what now?”/pm-status/pm-next
Finished a todo/pm-done TODO-xxx
Production / local errorpaste + /pm-fix
Before release/pm-all.pm/dashboard/
Need architecture diagrams/pm-arch or pm arch
Hand-off / laptop switch/pm-export
/pm-init → /pm-status → /pm-done
         ↘ /pm-fix (incidents)
         ↘ /pm-all  → .pm/dashboard/  (release gate + overview)

Target .pm/dashboard/ (after /pm-all or pm dashboard)

.pm/dashboard/
  index.html    # visual dashboard — open in a browser
  overview.md   # IDE tables: KPI, bars, module risk ranking, hot/todos
  findings.md   # aggregated open findings
  todos.md      # aggregated open todos
  stats.json    # counts + health_score + module_risk
  README.md     # same as overview.md (entry alias)

Repository Layout

pm-manager/
  pyproject.toml                 # uv/pip package (pm / pm-manager entrypoints)
  src/pm_manager_cli/            # Cross-platform CLI (incl. pm dashboard)
  skills/pm-manager/             # Installable Agent Skill (skills.sh)
    SKILL.md                     # Router skill
    AGENTS.md                    # Agent-oriented notes
    templates/commands/          # Slash/skill command prompts
    templates/pm/                # Files copied into target .pm/
    memory/ROUTING.md            # NL → command map
  scripts/python/                # Thin wrappers (prefer `pm init`)
  scripts/powershell/            # Legacy Windows helpers
  adapters/cursor/               # Cursor skill variants
  adapters/claude-code/          # Claude Code command files
  skills.sh.json                 # skills.sh groupings

Prerequisites

  • Windows / macOS / Linux
  • uv (recommended) for uv tool install
  • Python 3.11+
  • Git (so exclude rules can be written)
  • An AI coding agent that supports skills or project slash commands (Cursor or Claude Code recommended)
  • Optional: Spec Kit if you already use constitution/specs

Support

Questions and bugs: open a GitHub issue.

License

This project is licensed under the MIT License.

Frequently Asked Questions

What is pm-manager?

pm-manager is an open-source ai agents skill for AI coding assistants such as Claude Code, Codex CLI, and ChatGPT, built by wei63w. Know what to fix next — local .pm governance skill pack for AI coding agents (Spec Kit–inspired). It has 85 GitHub stars.

Is pm-manager safe to use?

Yes. pm-manager passed SkillsLLM's automated security scan — a dependency vulnerability audit plus prompt-injection heuristics — with no high-severity issues. You can read the full report in the Security Report section on this page.

How do I install pm-manager?

Clone the repository with "git clone https://github.com/wei63w/pm-manager" and add it to your Claude Code skills directory (see the Installation section above).

What programming language is pm-manager written in?

pm-manager is primarily written in Python. It is open-source under wei63w on GitHub, so you can review or fork the full source.

Are there alternatives to pm-manager?

Yes. SkillsLLM lists many other AI Agents skills you can browse and compare side by side. Open the AI Agents category from the badge at the top of this page, or use the Related Skills and comparison links further down to weigh pm-manager against similar tools.

Comments (0)

No comments yet. Be the first to share your thoughts!

ECC

by affaan-m

10

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

242,21936,702JavaScript
AI Agentsai-agentsanthropicclaude-code
View details
15

An agentic skills framework & software development methodology that works.

234,96620,863Shell
AI Agentsai-agentsbrainstorming
View details

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

185,94028,768JavaScript
AI Agentsai-agentsanthropicclaude-code
View details

cc-switch

by farion1231

3

A cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Grok Build & Hermes Agent. Only official website: ccswitch.io

128,8688,826Rust
AI Agentsclaude-codeai-tools
View details

claude-code

by anthropics

Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.

120,03119,897Shell
AI Agents
View details

Developers Also Liked

Based on votes and bookmarks from developers who liked this skill

ECC

by affaan-m

10

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

242,21936,702JavaScript
AI Agentsai-agentsanthropicclaude-code
View details
15

An agentic skills framework & software development methodology that works.

234,96620,863Shell
AI Agentsai-agentsbrainstorming
View details

n8n

by n8n-io

12

Fair-code workflow automation platform with native AI capabilities. Combine visual building with custom code, self-host or cloud, 400+ integrations.

201,88160,308TypeScript
MCP Serversapisai-tools
View details

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

185,94028,768JavaScript
AI Agentsai-agentsanthropicclaude-code
View details

cc-switch

by farion1231

3

A cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Grok Build & Hermes Agent. Only official website: ccswitch.io

128,8688,826Rust
AI Agentsclaude-codeai-tools
View details