fix: add max-ai-credits: 1500 to safe-output-health workflow

[Copilot]

The safe-output-health daily monitoring workflow was exhausting its AI budget, triggering the legacy effective-token rate-limit signal. The workflow had no max-ai-credits set, falling back to the compiled default of 1000 AICs — too low for a full 24h audit run.

Changes

• safe-output-health.md: Added max-ai-credits: 1500 to frontmatter, consistent with other daily monitoring workflows (schema-consistency-checker, daily-safe-output-optimizer, workflow-health-manager, etc.)
• safe-output-health.lock.yml: Recompiled — agent job apiProxy.maxAiCredits updated from 1000 → 1500

engine: claude
max-ai-credits: 1500  # added
tools:
  cli-proxy: true

[github-actions]

🧠 Matt Pocock Skills Reviewer has completed the skills-based review. ✅

[Copilot]

Pull request overview

Adjusts the safe-output-health daily monitoring workflow’s AI credit budget so a full 24h audit run can complete without hitting the compiled default limit.

Changes:

• Added max-ai-credits: 1500 to .github/workflows/safe-output-health.md frontmatter.
• Recompiled the workflow lockfile so apiProxy.maxAiCredits is updated from 1000 to 1500.

Show a summary per file

File	Description
.github/workflows/safe-output-health.md	Adds explicit max-ai-credits: 1500 budget to prevent running out of credits during daily audits.
.github/workflows/safe-output-health.lock.yml	Regenerated lockfile reflecting the new maxAiCredits value in the embedded AWF config.

Copilot's findings

Tip

Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

• Files reviewed: 2/2 changed files
• Comments generated: 0

[github-actions]

✅ PR Code Quality Reviewer completed the code quality review.

[github-actions]

🧪 Test Quality Sentinel completed test quality analysis.

No test files were added or modified in PR #37506. The PR only changes workflow configuration files (.github/workflows/safe-output-health.lock.yml and .github/workflows/safe-output-health.md). Test Quality Sentinel skipped — no tests to analyze.

[github-actions]

✅ Design Decision Gate 🏗️ completed the design decision gate check.

No ADR enforcement needed: PR #37506 does not have the 'implementation' label and has 0 new lines of code in business logic directories (≤100 threshold). Neither enforcement condition is met.

[github-actions]

Hey @Copilot 👋 — great catch on the safe-output-health workflow hitting its AI credit ceiling! Bumping max-ai-credits to 1500 and recompiling the lock file is exactly the right fix, and the description clearly explains the root cause and aligns it with the other daily monitoring workflows.

One small item worth considering before merge:

• No test coverage — because this is a workflow configuration change rather than source code, automated unit tests don't directly apply. However, it would be valuable to add or update any existing test fixtures or snapshot files that assert on the compiled lock file's apiProxy.maxAiCredits value, if such tests exist in the repo. This keeps the compiled output from drifting silently in future.

If you'd like a hand checking whether lock-file snapshot tests need updating, you can use this prompt:

Search the repository for any test files or snapshot fixtures that reference `safe-output-health.lock.yml` or assert on `apiProxy.maxAiCredits`. If such tests exist, update them to reflect the new value of 1500 so the test suite stays green.

Generated by ✅ Contribution Check · 366 AIC · ⌖ 13 AIC · ⊞ 24.3K · ◷

[github-actions]

Clean config bump — approved

The only change is max-ai-credits: 1000 → 1500 in the workflow frontmatter, with the lock file correctly recompiled: frontmatter_hash updated, body_hash unchanged (body untouched), and maxAiCredits in the embedded config aligned to 1500.

💡 Minor observation

No comment or PR description explains why 1000 credits was insufficient (e.g., health-check runs timing out). Worth a one-liner for future auditors — but not a blocker.

🔎 Code quality review by PR Code Quality Reviewer · sonnet46 46.8K · ⌖ 13.3 AIC

[github-actions]

Skills-Based Review 🧠

Applied /diagnose — this is a clean, targeted bug fix that correctly addresses the root cause. Approving.

📋 Assessment

Key Themes

• Root cause correctly addressed: The missing max-ai-credits frontmatter key was silently falling back to the 1000 AIC compiled default — too low for a full 24h audit run. Adding 1500 directly fixes the exhaustion.
• Value choice is well-motivated: Consistent with the established baseline for other daily monitoring workflows (schema-consistency-checker, daily-safe-output-optimizer, workflow-health-manager), which is explicitly called out in the PR description.
• Lock file properly maintained: frontmatter_hash is correctly rotated, body_hash is unchanged (workflow body untouched), and maxAiCredits is accurately updated from 1000 → 1500 in the compiled awf-config blob.

Positive Highlights

• ✅ Minimal, surgical change — no unrelated modifications
• ✅ PR description explains the observed failure mode and value rationale clearly
• ✅ Lock file kept in sync with the markdown source
• ✅ Change class is consistent with project conventions

🧠 Reviewed using Matt Pocock's skills by Matt Pocock Skills Reviewer · sonnet46 364.5K · 109.4 AIC · ⌖ 13.4 AIC