Pr To Green

---
name: pr-to-green
description: >
  Orchestrate a GitHub PR from first push to merge-ready in a single pass: run ci-fix-loop until
  all checks are green, detect active AI review bots (Qodo, CodeRabbit), run their fix loops in
  sequence until zero unresolved threads remain, then confirm the PR is merge-ready. Use when the
  user says "drive this PR to green", "make this merge-ready", "ship this branch", or wants CI and
  review comments handled end to end without babysitting each loop separately.
license: MIT
allowed-tools: Bash, Read, Write, Edit
compatibility: Codex, Claude Code, Cursor, GitHub Copilot, Windsurf, Kiro, and other Agent Skills compatible tools. Requires git and gh (GitHub CLI) authenticated, with GitHub Actions configured.
metadata:
  targets: [_source-only]
  author: Oleg Koval
  tags:
    - github
    - pr
    - ci
    - loop
    - qodoloop
    - coderabbitloop
    - ci-fix-loop
    - workflow
    - automation
source: weekly-pattern-learner
source_reason: "ci-fix-loop, coderabbitloop, and qodoloop all exist independently: no skill orchestrates them into a single end-to-end 'drive PR to green' pass"
source_date: "2026-07-31"
---
<!-- Generated by scripts/build-adapters.sh. Do not edit directly. -->

> 🤖 *Auto-generated by **weekly-pattern-learner** · ci-fix-loop, coderabbitloop, and qodoloop all exist independently: no skill orchestrates them into a single end-to-end 'drive PR to green' pass*

# PR to Green

Drive a GitHub PR from first push to fully merge-ready: CI checks green, every active AI
review bot addressed, zero unresolved threads.

## Inputs

- **PR number** (optional): detect from current branch if not given.
- **Max total iterations** (optional, default: 10): across all loops combined.
- **Skip bots** (optional): `--skip-qodo`, `--skip-coderabbit` to limit which review loops run.

## Phase 1: Ensure the PR exists

```bash
gh pr view --json number,headRefName,state 2>/dev/null
```

If no open PR exists, run `olko:pr-description-writer` first to create one, then continue.

## Phase 2: CI fix loop

Run `olko:ci-fix-loop` until all required checks pass, or a blocker needs a human.

Check current state first:

```bash
gh pr checks --json name,state,conclusion \
  --jq '.[] | select(.conclusion != "success" and .conclusion != "skipped")'
```

If all checks already pass, skip to Phase 3. Otherwise run `olko:ci-fix-loop`.

**Stop entirely if ci-fix-loop returns `blocked`**: do not run review loops while CI is red,
as bots may refuse to review a failing branch and their feedback will shift after fixes land.

## Phase 3: Detect active review bots

```bash
# Check if Qodo has posted on this PR
gh api repos/{owner}/{repo}/pulls/<PR_NUMBER>/reviews \
  --jq '[.[] | select(.user.login == "qodo-code-review[bot]")] | length'

# Check if CodeRabbit has posted
gh api repos/{owner}/{repo}/pulls/<PR_NUMBER>/reviews \
  --jq '[.[] | select(.user.login == "coderabbitai[bot]")] | length'
```

If neither has reviewed yet, nudge both and wait:

```bash
gh pr comment <PR_NUMBER> --body "@coderabbitai review"
gh pr comment <PR_NUMBER> --body "@qodo-code-review review"
```

Poll (every ~15s, timeout ~5min) for a review from either bot. If neither responds, proceed
without a review-bot loop and note it in the report.

## Phase 4: Review bot loops

Run each detected loop to completion before starting the next.

**Order:** qodoloop first (actionability-first bot, stricter on must-fix findings), then
coderabbitloop (thoroughness-first, may surface additional style or security hints after Qodo's
fixes change the diff context).

### Qodo

If Qodo posted at least one review: run `olko:qodoloop`.

It handles fetching unresolved threads, reading each Agent Prompt block, applying the fix,
replying to the thread, resolving it, pushing, and polling for the next Qodo review.

If qodoloop returns rate-limited, note it and continue to CodeRabbit without stopping.

### CodeRabbit

If CodeRabbit posted at least one review: run `olko:coderabbitloop`.

Same structure: fetch unresolved threads, read `🤖 Prompt for AI Agents` blocks, fix, reply,
resolve, push, wait for the next CodeRabbit review.

If coderabbitloop returns rate-limited, record it in the report and stop.

## Phase 5: Final verification

Confirm the PR is truly merge-ready before reporting success:

```bash
# Required checks all green?
gh pr checks --json name,state,conclusion \
  --jq '[.[] | select(.conclusion != "success" and .conclusion != "skipped")] | length'

# Any unresolved review threads?
gh api graphql -f query='
  query($owner: String!, $repo: String!, $number: Int!) {
    repository(owner: $owner, name: $repo) {
      pullRequest(number: $number) {
        reviewThreads(first: 100) {
          nodes { isResolved }
        }
      }
    }
  }
' -f owner="{owner}" -f repo="{repo}" -F number=<PR_NUMBER> \
  --jq '.data.repository.pullRequest.reviewThreads.nodes | map(select(.isResolved == false)) | length'
```

If either count > 0, identify what remains and report specifically: do not claim the PR is
green when it is not.

## Report

```
PR-to-green complete.
  CI:           ✓ all checks green (2 fix iterations)
  Qodo:         ✓ 4 findings fixed, 0 unresolved
  CodeRabbit:   ✓ 6 findings fixed, 0 unresolved
  Blockers:     none
  Status:       merge-ready ✓

  OR

  CI:           ✓ green
  Qodo:         ✓ 2 findings fixed
  CodeRabbit:   ⚠ rate-limited, 1 finding still open (re-run tomorrow)
  Status:       partial: CodeRabbit quota exhausted, human review needed
```

## Chaining

| Before this skill | After this skill |
|---|---|
| `olko:git-commit` + push | `gh pr merge` |
| `olko:pr-description-writer` | merge or stakeholder review |