sprint-fanout

Drafts per-repo sprint plans in parallel from a PRD using the configured review engine.

1|Updated Jun 13, 2026
One-click install
npx skills add https://github.com/apireno/agent-workflow-template --skill sprint-fanout-apireno
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: sprint-fanout
Source: https://github.com/apireno/agent-workflow-template/tree/main/.claude/skills/sprint-fanout
Command: npx skills add https://github.com/apireno/agent-workflow-template --skill sprint-fanout-apireno

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Coordinating sprint planning across multiple repositories is slow and inconsistent when done by hand — each repo needs its own scoped plan derived from the same PRD, and manually prompting per repo invites drift and missed scope. This Skill fans a single PRD or goal document out to every active repo in the fleet registry and produces a scoped sprint-plan.md per repo in one pass. ## Core Features & Use Cases - Parallel multi-repo drafting: Reads the fleet registry (.cto/projects.yaml), stages a per-repo prompt combining the PRD with each repo's CLAUDE.md, and drafts plans in parallel via the resolved review engine (subagent, gemini, kimi, codex, or claude-p). - Scoped, structured output: Each plan lands at the target repo's docs/sprints/sprint-NN/sprint-plan.md with required sections (Scope, Tasks, Test Plan, Acceptance Criteria, Dependencies, Risks), and repos with no applicable work are marked via a NO_SCOPE_FOR_THIS_REPO sentinel. - Dry-run and filtering: Supports --repos, --sprint, and --dry-run flags so you can preview staged plans before writing them into repos. - Use Case: A CTO with a PRD for a cross-repo feature runs the skill to get scoped sprint plans for all four dev repos at once, then reviews the synthesized summary of tasks, effort estimates, and cross-repo dependency flags. ## Quick Start Ask the agent to draft sprint plans across all active repos for the PRD at docs/prds/my-feature.md, optionally limiting to specific repos or a sprint number.

Frequently Asked Questions about sprint-fanout

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I draft sprint plans for multiple repos from one PRD?

Run the skill with the PRD path as the first argument, e.g. /sprint-fanout docs/prds/my-prd.md. It reads the fleet registry, stages a per-repo prompt combining the PRD with each repo's CLAUDE.md, and drafts a sprint-plan.md per active repo in parallel.

How do I limit sprint planning to specific repos or a sprint number?

Pass --repos repo1,repo2 to restrict the fanout to named repos from the registry, and --sprint NN to set the sprint number used in output paths. Add --dry-run to stage plans in a temp directory without writing them into the repos.

Which review engines can draft the sprint plans?

The engine is resolved by scripts/agentic/resolve-review-engine.sh: REVIEW_ENGINE env var, then a .review-engine file, defaulting to subagent. CLI engines (gemini, kimi, codex, claude-p) draft inline in parallel; subagent or handoff stages prompts for the CTO to draft via the Agent tool.

Why does sprint-fanout report no CTO home found?

The skill must locate a directory containing .cto/projects.yaml, checked via CTO_HOME, CTO_REPO, CLAUDE_PROJECT_DIR, walking up from the working directory, or ~/.cto/home. Fix it permanently by writing the CTO home path to ~/.cto/home.

What happens when a PRD does not apply to a repo?

The drafting prompt instructs the engine to output NO_SCOPE_FOR_THIS_REPO when the PRD has no work for that repo. If the sentinel appears in the first five lines of the plan, the repo is marked SKIPPED and no sprint-plan.md is written for it.