[Outcome Report] Workflow Health Report — 2026-06-22

[github-actions]

Workflow Health — 2026-06-22

Executive read: Overall acceptance rate is strong (100%), but 21 items are still pending with zero-touch rate at 0% — accepted items consistently need human engagement. Three workflows are stuck (stuck Daily workflows) with single pending items and no progression. Five workflows are underdefined with no observable follow-up signal. Data quality is acceptable at 5.9% fallback evaluations.

Workflow	Status	Lifecycle health	References
Matt Pocock Skills Reviewer	🟨🟨🟨🟩🟨🟨🟨🟨🟨🟩🟨🟨🟨🟨🟨⬜	🟡 in flight	🟩 #40696 · 🟩 #40695 · ⬜ #40698
PR Code Quality Reviewer	🟨🟩🟨⬜	🟡 in flight	🟩 #40696 · ⬜ #40698
PR Sous Chef	🟩🟩🟨🟩🟨	🟡 in flight	🟩 #40695 · 🟩 #40695 · 🟨 #40695 · 🟩 #40662 · 🟨 #40662
Semantic Function Refactoring	🟩🟨	🟡 in flight	🟩 #40531 · 🟨 #40701
Test Quality Sentinel	🟨🟩	🟡 in flight	🟨 #40696 · 🟩 #40696
Daily Documentation Healer	🟨	🔴 stuck	🟨 #40700
Daily Model Inventory Checker	🟨	🔴 stuck	🟨 #40704
Daily Sentrux Report	⬜	⚪ underdefined	⬜ #40703
Q	⬜	⚪ underdefined	⬜ #40702
Release	⬜	⚪ underdefined	⬜ v0.80.8

Legend:

• Status: 🟩 accepted · 🟥 rejected · 🟨 pending · ⬜ unknown
• Lifecycle health: 🟢 resolving · 🟡 in flight · 🟠 aging · 🔴 stuck · ⚪ underdefined
• References: one linked item per status emoji, in the same order as the Status column

🔴 Action Items

1. Stuck pending items — Daily Documentation Healer ([doc-healer] DDUw improvement: auto-expired not_planned doc issues bypass both DDUw and doc-healer #40700) and Daily Model Inventory Checker ([model-inventory] Model alias inventory update - 2026-06-22 #40704) are each pending with single items and no resolution. These workflows may need manual review or the evaluation model needs refinement. Check if they are expected to self-resolve or require explicit closure.

2. Underdefined workflows — Daily Sentrux Report, Q, and Release workflows produced outcomes with no observable follow-up signal (marked as unknown/exists_only). These workflows need clearer acceptance/rejection criteria or dedicated evaluators to classify their outcomes meaningfully.

3. Zero-touch rate at 0% — All 8 accepted items required human engagement. This indicates the agent outputs are not production-ready without edits. Consider reviewing prompts and output templates for these workflows to improve autonomy and reduce editing burden.

4. Data quality flag — 2 items (5.9% of 34) were evaluated using only a generic existence check (weak signal). Acceptable threshold; recommend targeted evaluators for create_check_run, create_pull_request_review_comment, and update_pull_request types to strengthen evidence.

5. High pending volume — 21 items (61.8% of total) are pending. Monitor over next 2 report cycles; if pending count does not decrease, workflows may be generating outputs faster than they can be resolved.

Detailed metrics, evidence quality, workflow counts, and trends

Outcome Scorecard — 2026-06-22

Metric	Value	Status
Acceptance rate	100.0%	🟢 >80%
Zero-touch rate	0.0%	🔴 <25%
Waste rate	0.0%	🟢 <10%
Median time to resolution	49m 18s	—
Accepted	8 / 34	—
— strong evidence	5	merged, completed, approved
— medium evidence	3	engaged, retained
— weak evidence	0	existence only
Rejected	0	—
Ignored	0	no observable follow-up
Zero-touch	0 / 8	—
Pending	21	—
Runs checked	14	—

Per-Workflow Breakdown

Workflow	Accepted	Rejected	Ignored	Pending	Acceptance	Zero-touch
Matt Pocock Skills Reviewer	2	0	0	13	100.0%	0.0%
PR Code Quality Reviewer	1	0	0	2	100.0%	0.0%
PR Sous Chef	3	0	0	2	100.0%	0.0%
Semantic Function Refactoring	1	0	0	1	100.0%	0.0%
Test Quality Sentinel	1	0	0	1	100.0%	0.0%
Daily Documentation Healer	0	0	0	1	0.0%	0.0%
Daily Model Inventory Checker	0	0	0	1	0.0%	0.0%
Daily Sentrux Report	0	0	0	0	0.0%	0.0%
Q	0	0	0	0	0.0%	0.0%
Release	0	0	0	0	0.0%	0.0%

Evidence Quality

⚠️ 2 item(s) were evaluated using only a generic existence check (signal: target_exists_only). These contribute to accepted_weak and may overstate acceptance. Dedicated evaluators for add_reviewer, submit_pull_request_review, update_issue, update_pull_request, and other types provide stronger evidence.

📊 Measured by Outcome Collector · 23.4 AIC · ⌖ 3.61 AIC · ⊞ 6.9K

• expires on Jun 28, 2026, 5:07 PM UTC-08:00