pm-claude-skills/templates/pm-sprint-agent/README.md

PM Sprint Agent — Agent Template

An end-to-end sprint planning agent built from existing skills in pm-claude-skills. Pulls your backlog, calculates capacity, drafts the sprint plan, flags risks, and posts the result.

This is the first agent template in the pm-claude-skills library. It follows the architecture Anthropic introduced for financial services agent templates on May 5, 2026 — packaging skills + connectors + subagents into a single runnable workflow.

---

What it does

You point this agent at your team's backlog and a sprint goal. It does the rest:

1. Pulls open issues from Linear or Jira, filtered by the sprint scope
2. Calculates team capacity for the upcoming sprint (using the Capacity Analyst subagent)
3. Drafts a sprint plan using the sprint-planning skill from this library
4. Generates a sprint kickoff brief using the sprint-brief skill
5. Scores delivery risks for the proposed plan (using the Risk Scorer subagent)
6. Posts the result to a Slack channel for team review

End-to-end: roughly 90 seconds for a 25-issue backlog.

---

What's inside this template

templates/pm-sprint-agent/
├── README.md                          ← you are here
├── AGENT.md                           ← agent definition (system prompt + tool list)
├── orchestrate.sh                     ← orchestration script
├── skills/                            ← skills used by this agent (linked from main library)
│   ├── sprint-planning/SKILL.md       ← (symlink to ../../skills/sprint-planning/)
│   ├── sprint-brief/SKILL.md          ← (symlink to ../../skills/sprint-brief/)
│   ├── retro/SKILL.md                 ← (symlink to ../../skills/retro/)
│   └── project-status-report/SKILL.md ← (symlink to ../../skills/project-status-report/)
├── subagents/
│   ├── capacity-analyst.md            ← team capacity calculation subagent
│   └── risk-scorer.md                 ← delivery risk scoring subagent
├── connectors/
│   ├── README.md                      ← connector setup guide
│   ├── linear.example.json            ← Linear connector example config
│   └── jira.example.json              ← Jira connector example config
├── examples/
│   ├── input-example.md               ← what you feed the agent
│   └── output-example.md              ← what the agent produces
└── tests/
    └── smoke-test.md                  ← manual smoke test for new installations

---

Quick install (5 minutes)

Prerequisites

• Claude Code installed
• The full skills library installed: /plugin marketplace add mohitagw15856/pm-claude-skills
• A Linear or Jira workspace
• A Slack workspace (optional — for the post-to-Slack step)

Setup steps

1. Configure your connectors.

Open connectors/linear.example.json (or jira.example.json if you use Jira) and fill in your team's specifics. Save as connectors/linear.json (without the .example). Add your API token to the credentials section.

cd templates/pm-sprint-agent/connectors
cp linear.example.json linear.json
# Edit linear.json with your team_id, workspace_url, and API token

2. Configure the agent.

Open AGENT.md and update the configuration block at the top with your team's defaults — sprint length, capacity buffer, default Slack channel.

3. Test the smoke test.

Run the smoke test to verify everything is wired up:

bash orchestrate.sh --dry-run --sprint-goal "Test sprint planning"

If the dry-run completes without errors, you're ready to run a real sprint plan.

---

Running the agent

Standard usage

bash orchestrate.sh \
  --sprint-goal "Reduce checkout abandonment by 20%" \
  --sprint-number 23 \
  --team-size 5 \
  --duration-weeks 2

The agent will:

1. Pull open issues tagged with the sprint goal scope
2. Run the Capacity Analyst subagent to calculate available capacity
3. Run the sprint-planning skill to draft the sprint plan
4. Run the Risk Scorer subagent to flag delivery risks
5. Run the sprint-brief skill to generate the kickoff brief
6. Output everything to output/sprint-23-plan.md
7. (Optionally) post a summary to your configured Slack channel

Configuration options

Flag	Required	Default	Description
--sprint-goal	Yes	—	Short description of what the sprint should achieve
--sprint-number	Yes	—	Which sprint this is (e.g., 23)
--team-size	No	5	Number of engineers on the team
--duration-weeks	No	2	Sprint length in weeks
--capacity-buffer	No	0.2	Buffer for unplanned work (0-1 range)
--include-bugs	No	true	Include open bugs in the sprint plan
--post-to-slack	No	true	Post summary to Slack
--dry-run	No	false	Validate config without running the workflow

---

Why this architecture

The template follows Anthropic's three-component pattern:

Skills provide the structured output formats. The sprint-planning skill knows what a sprint plan should contain. The sprint-brief skill knows what a kickoff brief should look like. These already exist in this library — the agent doesn't reinvent them.

Connectors provide governed access to data. The agent doesn't hold your Linear API token in a prompt — it uses the configured Linear connector with proper authentication and rate limiting.

Subagents handle specialised analysis. Calculating team capacity isn't a one-shot generation task — it requires reading PTO calendars, assessing historical velocity, and adjusting for known capacity hits. That's a focused job for a subagent. Same logic for risk scoring.

This separation matters because each component can be tested, swapped, and improved independently. If you want to use a different sprint planning skill, swap it. If you switch from Linear to Jira, swap the connector. If you build a better capacity model, replace the subagent. The orchestration script doesn't change.

---

Customisation

Use your team's templates

This agent uses the generic sprint-planning and sprint-brief skills from the main library. If your team has specific conventions — story point scale, definition of done format, retro categories — fork the skills into the skills/ folder of this template and modify them. The orchestration script will pick up the local versions.

Add additional analysis steps

Add new subagents to subagents/ for any specialised analysis your team needs — engineering manager reviews, stakeholder impact assessment, dependency mapping. Update orchestrate.sh to call them at the appropriate point in the workflow.

Switch ticketing systems

The connectors are decoupled from the orchestration. Swap linear.json for jira.json (or build a connector for any other system) without touching the agent definition.

---

Limitations and honest caveats

Capacity calculation is heuristic, not exact. The Capacity Analyst subagent makes reasonable estimates based on historical velocity and team size. For more accurate capacity calculation, integrate with your team's actual time-tracking system.

Risk scoring is directional, not predictive. The Risk Scorer flags items that historically correlate with delivery risk (large story points, dependencies, team members on PTO). It doesn't predict what will actually slip. Use it as a discussion starter, not a forecast.

Linear and Jira are tier-1 supported. Other ticketing systems (Shortcut, Asana, Trello, ClickUp) can be added by following the connector pattern in connectors/README.md but aren't included out of the box.

No autonomous execution. This template runs as a Claude Code plugin — meaning it produces outputs for human review, it doesn't autonomously create or modify tickets. For autonomous execution, deploy via Claude Managed Agents using the same skills, connectors, and subagent definitions.

---

Contributing

If you build on this template — adding a new connector, improving the subagents, supporting a new ticketing system — consider raising a PR back to the main repo. Improvements that benefit the broader community are welcome.

For a full template contribution guide, see templates/CONTRIBUTING.md (coming soon).

---

Where to learn more

• Anthropic's announcement of agent templates (May 2026)
• Anthropic's Claude Managed Agents documentation
• The pm-claude-skills main README
• Part 16 article — Building My First Agent Template (link added when published)

---

Built and maintained by Mohit Aggarwal | First agent template in pm-claude-skills