feat: implement issue #195 — bug(dev-lead): engine.sh defines multi-tier model pipeline but all tasks always invoke ENGINE_ACTION_MODEL (sonnet-4-6)

[don-petry]

Closes #195

Implemented by dev-lead agent. Please review.

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.

[coderabbitai]

Warning

Review limit reached

@don-petry, we couldn't start this review because you've reached your PR review rate limit.

More reviews will be available in 35 minutes and 39 seconds. Learn how PR review limits work.

Your organization has run out of usage credits. Purchase more in the billing tab.

⌛ How to resolve this issue?

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.

We recommend that you space out your commits to avoid hitting the rate limit.

🚦 How do rate limits work?

CodeRabbit enforces hourly rate limits for each developer per organization.

Our paid plans include higher PR review limits than trial, open-source, and free plans. In all cases, reviews become available again over time. During sustained high-volume PR review activity, CodeRabbit may temporarily slow when the next review becomes available.

Please see our Fair Usage Limits Policy for further information.

ℹ️ Review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: ASSERTIVE

Plan: Pro

Run ID: 33954dcf-bbca-4698-b3cd-20416b52ac36

📥 Commits

Reviewing files that changed from the base of the PR and between 28c0731 and 631695e.

📒 Files selected for processing (6)

• scripts/dev-lead-fix-ci.sh
• scripts/dev-lead-fix-issue.sh
• scripts/dev-lead-fix-reviews.sh
• scripts/engine.sh
• tests/dev-lead/unit/test_model_dispatch.bats
• tests/dev-lead/unit/test_provider_headroom.bats

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch dev-lead/issue-195-20260610-2113

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

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

[gemini-code-assist]

Code Review

This pull request introduces intent-based model dispatching and provider headroom checks to optimize engine selection and prevent rate-limiting issues. It adds helper functions model_for_intent and check_provider_headroom in scripts/engine.sh, updates the dev-lead scripts to pass intent types, and includes comprehensive unit tests. The review feedback highlights several robustness improvements in scripts/engine.sh, including validating that rate-limit values are numeric to prevent crashes under set -eu, avoiding the curl -sI -X POST anti-pattern, and parsing HTTP headers more reliably by handling potential whitespace variations.

[don-petry]

Dev-Lead — review-changes (applied)

Changes committed and pushed.

[Copilot]

Pull request overview

This PR addresses issue #195 by making the dev-lead “writer” path actually use the multi-tier model configuration in engine.sh, and by improving provider fallback behavior with a proactive headroom check before invoking an engine.

Changes:

• Add intent-based model selection (model_for_intent) and plumb intent through run_writer_with_fallback, so different dev-lead tasks can use different tier models.
• Add check_provider_headroom and integrate it into the fallback loop to proactively skip engines near quota exhaustion.
• Add new Bats unit tests covering model dispatch and provider headroom behavior.

Reviewed changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
scripts/engine.sh	Adds intent→model mapping and provider headroom probing; updates run_writer_with_fallback to select tier model per intent and engine.
scripts/dev-lead-fix-reviews.sh	Passes INTENT_TYPE into run_writer_with_fallback so model selection matches the invoked intent.
scripts/dev-lead-fix-issue.sh	Passes explicit fix-issue intent into run_writer_with_fallback.
scripts/dev-lead-fix-ci.sh	Passes explicit fix-ci intent into run_writer_with_fallback.
tests/dev-lead/unit/test_provider_headroom.bats	Adds unit tests for check_provider_headroom and for headroom-driven fallback skipping.
tests/dev-lead/unit/test_model_dispatch.bats	Adds unit tests for model_for_intent and intent-based model selection through run_writer_with_fallback.

[don-petry]

Dev-Lead — review-changes (applied)

Changes committed and pushed.

[don-petry]

Dev-Lead — waiting on PR blockers (intent: fix-reviews)

PR: #556
No changes were committed, but the PR still has blocking checks or reviews (failing or cancelled checks, or changes-requested reviews). The retry cron will re-attempt automatically. Next attempt after: 2026-06-10T23:40:47Z

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[donpetry-bot]

Automated review — APPROVED ✓

Risk: LOW
Reviewed commit: 631695e20026a56d48c8e22f0223fc7c179f4538
Review mode: triage-approved (single reviewer)

Summary

Implements issue #195 (multi-tier model dispatch + proactive provider headroom checks). The dev-lead writer pipeline now actually uses the tiered model definitions in engine.sh — human-pr/fix-bot-comment route to haiku, fix-reviews/fix-ci/rebase to sonnet (action), fix-issue/human to deep — and a fail-open check_provider_headroom skips engines at/above 75% usage (configurable via DEV_LEAD_USAGE_THRESHOLD). The three dead-code-in-dev-lead-context functions (run_triage, run_agentic, run_duck) are now documented as review-one-pr.sh-only rather than removed, which is a reasonable scope choice. +571/-8 across 6 files, including 463 lines of new bats coverage.

Linked issue analysis

Closes #195. All acceptance criteria addressed:

• Model dispatch: model_for_intent matches the issue's spec exactly; all three caller scripts (dev-lead-fix-ci.sh, dev-lead-fix-issue.sh, dev-lead-fix-reviews.sh) pass an intent string through.
• Failover (bugs 1 & 2): Already resolved upstream in the existing run_writer (output is captured via tee and is_rate_limited_files; copilot path uses copilot_chat ... --yolo directly rather than falling back to claude).
• Proactive usage monitoring (bug 3): check_provider_headroom probes Anthropic and GitHub Models rate-limit headers, fail-opens on probe failure, and is wired in ahead of each engine attempt in the fallback loop.
• Dead code: run_triage/run_agentic/run_duck annotated as review-one-pr.sh-only (documented rather than removed — acceptable).

Findings

No blocking issues.

Minor observations (non-blocking):

• The claude headroom probe sends a real POST /v1/messages with max_tokens:1, so each writer invocation burns ~1 token even on the happy path. This is consistent with the issue spec's intent ("barely consumes quota") and is bounded.
• Copilot probe correctly guards against unset/placeholder tokens before making a network call — good defensive choice.
• bats coverage is thorough (16 dispatch tests + 13 headroom tests including threshold-edge and fail-open cases).

CI status

All required checks green: Lint, ShellCheck, bats, validate-fixtures, validate-agent-profiles, unit-tests, Compile agentic workflows, Agent Security Scan, CodeQL (actions + python), Secret scan (gitleaks), SonarCloud (Quality Gate passed), AgentShield, dev-lead dispatch. coderabbitai approved; gemini-code-assist and copilot-pull-request-reviewer commented without blocking. Skipped checks (dependabot, language-specific dependency audits) are expected — not applicable to this PR.

---

Reviewed automatically by the PR-review agent (single-reviewer mode: opus 4.7). Reply if you need a human review.