feat(ci): add exact cross-repository scheduler target

[seonghobae]

Summary

• add a bounded merge-scheduler entrypoint for one exact target_repository + pr_number
• validate the canonical repository against the existing central allowlist and re-read live open same-repository PR metadata before using cross-repository credentials
• bind concurrency to target repository and PR, resolve the live base branch, and route scheduler read/actions credentials to the target while retaining the central runner token for trusted review dispatch
• document the evidence-only dispatch payload and keep merge/branch mutation opt-in

Incident evidence

After central lock repair #639, a direct OpenCode retry for Naruon #1179 correctly failed actor validation because only github-actions[bot] may create privileged review dispatches. Required workflows in target repositories cannot be manually dispatched or rerun through the Actions API, while the only central fallback was an unbounded organization sweep. This left no safe way to validate one current head through the trusted scheduler.

Validation

• 659 full pytest tests passed
• scripts/ci statement coverage: 100% (5,599 statements)
• interrogate docstring coverage: 100%
• focused scheduler/agent/shell contracts: 90 passed
• executable accepted, unallowlisted, and cross-repository fake GitHub API cases passed
• actionlint passed for the changed scheduler workflow
• repository-wide Ruff passed
• CodeGraph initialized, synchronized, and used for impact inspection

[coderabbitai]

Warning

Review limit reached

@seonghobae, you've reached your PR review limit, so we couldn't start this review.

Next review available in: 15 minutes

Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available.
You're only billed for reviews past your plan's rate limits ($0.25/file).

How can I continue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews.

How do review limits work?

CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 3fa156dc-e88c-4f70-8efb-135677613434

📥 Commits

Reviewing files that changed from the base of the PR and between 488ad72 and 4131d84.

📒 Files selected for processing (5)

• .github/workflows/pr-review-merge-scheduler.yml
• docs/org-required-workflow-rollout.md
• tests/test_opencode_agent_contract.py
• tests/test_opencode_workflow_shell_syntax.py
• tests/test_required_workflow_queue_contract.py

---

_{Comment @coderabbitai help to get the list of available commands.}

[seonghobae]

Controlled bootstrap merge evidence for current head:

• Exact head: 4131d848535779094ca4beb1150a2fca055a5e33
• Exact base: 488ad725148453a3d85d6978c3d4763f30875ca5
• Local evidence: 659 tests passed; statement coverage 100% across 5,599 scripts/ci statements; docstrings 100%; Ruff and changed-workflow actionlint passed.
• Live evidence: coverage-evidence, CodeQL (Actions/Python), dependency-review, OSV, Trivy, Semgrep, Bandit, gitleaks, SBOM, and all non-provider required checks passed; current-head medium-or-higher code-scanning alerts and open Dependabot alerts are empty; unresolved review threads are zero.
• External reviewer evidence: OpenCode run 30428210073 exhausted all configured providers (GitHub Models permission failure, OpenAI quota exhaustion, OpenRouter credit exhaustion). Strix also recorded backend unavailability without a vulnerability artifact. Noema is correctly fail-closed behind a primary OpenCode approval.
• Exception: temporarily disable only ruleset 17921150 enforcement and administrator branch-protection enforcement, merge with the exact-head guard, then restore both immediately. The required checks, review counts, and conversation policy remain configured and are not rewritten.

This exception is limited to bootstrapping the exact-target central scheduler needed to re-run trusted current-head reviews across repositories; provider outage is not treated as a source-code blocker.