sprint-cycle-router

Routes cron-triggered sprint development cycles through a decision tree of GitHub automation branches.

Updated Aug 17, 2026
One-click install
npx skills add https://github.com/kai-kou/gem-hunter --skill sprint-cycle-router-kai-kou
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: sprint-cycle-router
Source: https://github.com/kai-kou/gem-hunter/tree/main/.claude/skills/sprint-cycle-router
Command: npx skills add https://github.com/kai-kou/gem-hunter --skill sprint-cycle-router-kai-kou

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Unattended sprint development requires deciding, on every scheduled run, which of many possible tasks (PR review follow-up, spec sync, backlog grooming, new sprint kickoff, hygiene audits) should execute first. This Skill provides a single decision-tree router that evaluates repository state via GitHub Issues, PRs, and labels, then executes exactly one branch per run so a cron routine can advance development autonomously. ## Core Features & Use Cases - Decision-tree routing: Evaluates 10 prioritized branches per firing (breaking-change response, own-PR collection, in-progress sprint resume, SP-to-Issue sync, new sprint kickoff with TDD, improvement issue digestion, retro-try handling, hygiene, weekly refinement, spec-sync verification) and runs only the first matching one. - API channel detection: Re-determines every run whether to use GitHub MCP tools, the gh CLI, or curl REST fallback, with fail-closed no-op behavior when all channels are unavailable. - Production drift inspection: Detects when main HEAD diverges from the deployed Cloudflare Workers build, re-triggers builds, and escalates via Slack with threshold-based notification suppression. - Use Case: Configure a 2-hour cron routine; each firing the router checks for open PRs needing review, stale in-progress issues, or a ready SP-n issue, then either resumes interrupted work or starts the next sprint without human intervention. ## Quick Start Ask the AI to run the sprint self-driving routine by saying "run the sprint cycle router and advance development for this repository".

Frequently Asked Questions about sprint-cycle-router

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

FAQPage Schema
How do I automate sprint development with a cron routine?▼

Configure a single recurring routine that invokes this router every N hours. Each firing evaluates repository state through GitHub Issues and PRs, picks the first matching branch from the decision tree, and executes only that one task, keeping per-run cost low.

How does the router decide which task to run each cycle?▼

It checks branches in priority order: breaking-change spec sync first, then own PR review follow-up, stale in-progress sprint resume, SP-to-Issue sync, new sprint kickoff, improvement and retro-try issue digestion, hygiene, and weekly refinement. Only the first matching branch executes.

Does it work when the GitHub MCP tools or gh CLI are unavailable?▼

Yes. Every firing re-detects the API channel in order: GitHub MCP tools, then gh CLI, then curl REST with a token. If all channels fail, it performs a safe no-op rather than creating partial local state.

How does it detect production deployment drift on Cloudflare Workers?▼

Each firing runs a drift check comparing main HEAD against the deployed build. If drift exists it re-triggers Workers Builds; if the trigger is unconfigured or judgment fails three consecutive times, it escalates to Slack with mention-based notification.

What are the limitations of unattended sprint automation?▼

State persists only through GitHub artifacts like labels, comments, and branches, since each firing is ephemeral. It intentionally ignores other people's manual PRs, and blocked or waiting-user labeled items are excluded to prevent infinite fix loops.