Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+22 more

    Issue to pr Runner

    1

    Turns a plain-language prompt into a GitHub issue, branch, isolated-worktree implementation, tests, reviewed PR and a confirmed merge, using 8 specialized sub-agents.

    $14

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    issue-to-pr-runner

    issue-to-pr-runner

    Example session with this skill installed

    run issue-runner on: add an optional "location" field to Event so organizers can say where it takes place. Show it on the event detail page and accept it in the creation form.

    • Read your context and instructions
    • Compiled the issue-to-pr-runner

    Pipeline complete: issue #42 → PR #43, ready to merge (waiting for your confirmation).

    SUMMARY
    Classified as NEW_ISSUE (single feature, no matching open issue). Risk level: medium (public API contract changes). I paused once to confirm "location" should be optional; you chose optional. 9 phases ran, 0 retries needed.

    DELIVERABLE

    PhaseAgentResult
    Intentintent-classifierNEW_ISSUE, confidence 0.94
    Specprompt-optimizer5 acceptance criteria, complexity: small
    Riskrisk-analyzermedium: 1 contract change, 1 test gap
    Setuporchestratorissue #42, branch runner/issue-42-add-location-to-event
    Codeimplementer (worktree)5 files modified, 1 migration created
    Reviewregression-checkerpass, 0 major findings
    Teststest-writer6 tests added, 38/38 passing (pnpm test)
    PRorchestrator#43, "Closes #42"
    Reviewpr-reviewerapprove, CI green, mergeable

    Files changed

    • prisma/schema.prisma: Event.location String? (nullable)
    • prisma/migrations/…_add_event_location/migration.sql
    • src/events/dto/create-event.dto.ts: optional, max 200 chars
    • src/events/events.service.ts: persists location
    • web/components/EventForm.tsx: new "Location" input
    • web/pages/events/[id].tsx: shows location when set
    • src/events/events.service.spec.ts: create with/without location, read-back, length validation

    Decisions recorded in the PR

    • Field is nullable so already-deployed clients keep working (mitigation from risk-analyzer).
    • No change to the list view, which was out of scope.

    NEXT STEPS

    1. Merge PR #43 with squash (recommended): reply "merge" or choose "View diff first".
    2. Deploy the migration before the new web build.
    3. Follow-up worth its own issue: show location in the events list and filters.

    State saved in .claude/runner-state/issue-42/state.json, so the run can be resumed at any time.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    From a one-line prompt to a reviewed pull request

    issue-to-pr-runner is a Claude Code skill that runs the whole path from "I want X" to a merged PR. Describe the work in plain language. It creates the GitHub issue, opens a dedicated branch, implements the change in an isolated git worktree, writes the tests, opens a PR, reviews it, and asks you before merging.

    Eight specialized sub-agents each do one job: they classify your intent, turn the prompt into a clear spec, assess risk, implement, check for regressions, write tests, and review the PR. Each step passes a structured result to the next.

    What you get

    • Prompt → issue → branch → PR in one run, with the issue linked to the PR (Closes #N).
    • Risk analysis before any code is written. High-risk changes (data loss, public API changes) stop and ask you first.
    • Isolated worktree. The implementation never touches your main checkout.
    • Regression check and tests. A static review of the diff, then new tests in your existing framework.
    • Independent PR review with a clear verdict and at most a handful of actionable comments.
    • Parallel features. A prompt with several independent tasks becomes several pipelines, each with its own issue, branch and PR.
    • Resumable. State is saved after every phase, so an interrupted run picks up where it stopped.

    You stay in control

    • Nothing is merged without your explicit confirmation.
    • The implementer agent never commits or pushes; the orchestrator does, on a dedicated branch.
    • Retries are capped (2 per phase by default) and failures are escalated to you with full context.

    Works on any stack

    Detects the test command automatically for npm, pnpm, yarn, bun, pytest, cargo, go, flutter, rspec, maven, gradle and dotnet. Prefer your own? Set testCommand in .claude/issue-runner.config.json.

    Requirements

    • Claude Code (uses sub-agents and worktree isolation)
    • node and an authenticated gh CLI on PATH
    • No npm dependencies

    Good to know

    • Cost depends on repo size and task complexity. The pipeline uses lighter models for classification and spec writing, and a stronger one for implementation and review.
    • Works with prompts in any language.
    • Commercial license for individual use on personal and client projects (see LICENSE.txt).

    How to use

    Skills on Agensi run inside your AI agent. Copy the prompt, paste it into your agent, and it connects to Agensi and installs this skill for you. If a payment is needed, your agent shows it to you first.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill passes our trust checks below.

    Security scanned

    Passed our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scanned

    Verified clean 4 days ago

    • Passed all security checks, Safe to install

    Requires

    Git

    Listed4 days ago

    What's inside

    Frequently Asked Questions