kurone-kito/idd-skill

Turn AI coding from one-off prompts into a GitHub-native delivery loop
that keeps going until the issue is closed.
IDD Skill gives a repository a portable Issue-Driven Development
workflow: agents discover ready issues, claim ownership, implement in a
branch, open a PR, handle review feedback, wait for CI, merge under the
policy you select, and clean up. The whole loop ships as plain Markdown
instruction files that live in your repository — yours to inspect,
fork, and audit.
Why IDD exists
A capable coding agent should finish work the way a strong teammate
does: visibly, verifiably, and all the way to done. Most AI coding
breaks down not at writing code but at everything around it:
- Two agents can pick the same issue.
- Review comments can be accepted but never fixed.
- CI can finish after the agent has already stopped watching.
- A PR can merge while fresh review activity is still arriving.
- The workflow rules can hide inside a vendor platform instead of your
repository.
IDD turns those moving parts into one auditable GitHub-native loop.
Issue comments record claims, review snapshots, decisions, holds, and
cleanup markers — so any agent can resume any work without guessing.
What you get
- **Collision-resistant parallel work** — claim and heartbeat markers
make issue ownership visible without a central coordinator.
- **Fewer stalled PRs** — the loop keeps watching review comments, CI,
and follow-up fixes after the pull request opens.
- **Portable rules, not a black box** — plain Markdown you can
inspect, fork, and customize in your own repository.
- **Agent choice** — works across GitHub Copilot, Claude Code, OpenAI
Codex CLI, OpenCode, Grok Build, Cursor CLI, and Antigravity CLI
(formerly Gemini CLI);
review policy profiles let you
swap the default Copilot advisory gate.
- **No service to operate** — no server, scheduler, or additional SaaS
account; import the template files and start.
Proven in production
IDD is not a demo. It is the workflow this repository has been built
with since May 2026:
- **3,200+ issues** turned into merged pull requests across private
work repositories running IDD.
- **1,700+ pull requests** merged in this public repository alone,
through multi-agent bursts of x4-6 parallel sessions (x8-10 in the
originating private deployment).
- Early August 2026 peaked at **600+ commits** and **150+ pull
requests** in a single day across the private work repositories
and this public repository.
- **Not zero-failure by design** — edge cases found in production
come back as issues, and the loop fixes itself.
_As of 2026-10._
Token-cost measurement is in progress; see
docs/token-cost.md for the methodology.
Quick Start
Open an AI-agent session in the repository that should adopt IDD and
say:
> Import and onboard github:kurone-kito/idd-skill's IDD into this
> repository.
If the agent needs an explicit URL, use:
> Read
> https://raw.githubusercontent.com/kurone-kito/idd-skill/main/idd-template/ONBOARDING.md
> and onboard this repository for Issue-Driven Development.
Then start the loop:
> Start the IDD workflow in this repository.
The agent discovers a ready issue, claims it, and follows the loop
through work, PR review, CI, the selected merge policy, and cleanup.
See Getting started for the full first-run
path, including the optional IDD doctor validation.
Prerequisites
The full loop needs git, an authenticated gh CLI (or an equivalent
GitHub MCP integration), jq, a REST client such as curl, and