Project Management — Domain Orchestrator & Delivery Loop
This skill orchestrates project-delivery work across eight project-management sub-skills (sprint/velocity analytics, portfolio health, Jira/JQL, Confluence, Atlassian admin, templates, meeting analysis, team comms). It routes inquiries to the correct sub-skill via a deterministic signal router and can drive a full goal-to-close delivery loop with live Jira data, machine-run verification gates, and governance rules. It produces digests, flow metrics, and delivery-loop artifacts.
✨ What it does
- Routes PM inquiries to one of eight sub-skills using a deterministic signal router script.
- Bridges live Jira snapshots into domain analytics tools for flow metrics, sprint velocity, and Monte Carlo forecasts.
- Enforces a delivery-loop contract with machine-run verification gates and terminal states.
- Enforces governance rules: human ownership, machine-checkable acceptance, reversible-first writes, and forecast ranges.
- Provides a forcing-question library to lock lane-defining decisions before execution.
- Generates digests with canon-cited challenges and output artifacts like flow_metrics.json and plan.json.
🎯 When to use it
- When you need to coordinate delivery work across multiple PM sub-skills, such as sprint health, portfolio reporting, or Jira configuration.
- When you want to run a bounded agentic delivery loop from goal to verified close, pulling live Jira data and enforcing governance.
- When you need to route a PM inquiry to the correct specialized skill without guessing.
- When you need to produce flow metrics, velocity forecasts, or portfolio health reports from live Jira data.
🚀 How to use
Trigger by describing a PM goal or symptom, e.g., 'our sprints feel off', 'project health report', 'audit our Jira permissions', 'when will it be done', or 'run the delivery loop'. The skill will route to a sub-skill or start a delivery loop. Provide the goal text or inquiry; for loops, expect to supply a clear deliverable and acceptance criteria. Example prompts:
Run the delivery loop to get sprint 14 to a verified close.
Produce a portfolio health report from live Jira data.
📄 Output: Depending on mode, outputs include a sub-skill artifact with a ≤200-word digest, a flow_metrics.json report, or .agent-harness/plan.json with state and gate verdicts.
📦 Add this skill to Claude Code
# 1. Get the skills repo
git clone --depth 1 https://github.com/alirezarezvani/claude-skills /tmp/claude-skills
# 2. Copy this skill into your project (or ~/.claude/skills for all projects)
mkdir -p .claude/skills
cp -r /tmp/claude-skills/project-management/skills/pm-skills .claude/skills/pm-skillsSkill source: project-management/skills/pm-skills/SKILL.md
❓ FAQ
How does the skill decide which sub-skill to use?
It runs a deterministic signal router script (pm_goal_router.py) that classifies the inquiry and names the sub-skill; if ambiguous, it asks one clarifying question.
Can I get a forecast for when work will be done?
Yes, using Monte Carlo simulation on completed work items, but it refuses if fewer than 10 items are completed because thin history forecasts are unreliable.
What happens if the delivery loop exceeds its budget?
An exhausted budget is an escalation, not a success; the loop stops and escalates to the named human with an evidence log.
🤖 Overview, features, install steps and FAQ were generated from the project's SKILL.md on Sep 4, 2026. Always check the original source before running commands.