[daily regulatory] Regulatory Report - 2026-03-29

[github-actions[bot]]

8 daily report discussions were reviewed for 2026-03-29. Data quality is generally good across operational metrics. The most significant findings are: (1) 21 workflow compile failures in the Static Analysis report due to a shared missing import — a systemic issue affecting 11.8% of workflows; (2) a high run failure rate of 41% (16/39 completed runs) in the Safe Output Health report, though per-item success remains healthy at ~94%; and (3) a PR count discrepancy between the Team Evolution and Copilot Agent Analysis reports requiring scope clarification. Five report types present on 2026-03-27 are absent today (Observability, Performance, Code Metrics, DeepReport, Prompt Clustering), which is consistent with the 2026-03-28 pattern.

The security hardening sprint is proceeding well — the DIFC integrity filter is actively working (221 events, 26 external contributors filtered), the firewall maintained a healthy 1.3% block rate, and the Copilot agent merge success rate improved to 79.3% (+4.3% over 3 days). No critical regulatory violations were identified.

📋 Full Regulatory Report

📊 Reports Reviewed

#	Report	Discussion	Created	Status
1	Sergo Error Patterns Analysis	#23480	2026-03-29T20:25Z	✅ Valid
2	DIFC Integrity-Filtered Events	#23479	2026-03-29T20:20Z	✅ Valid
3	Static Analysis Report	#23478	2026-03-29T19:44Z	⚠️ Issues (compile failures)
4	Safe Output Health Report	#23459	2026-03-29T13:20Z	⚠️ Issues (high run failure rate)
5	Copilot Agent Session Insights	#23444	2026-03-29T12:01Z	✅ Valid
6	Daily Firewall Report	#23441	2026-03-29T11:52Z	✅ Valid
7	Copilot Agent Analysis	#23438	2026-03-29T11:33Z	✅ Valid
8	Daily Team Evolution Insights	#23427	2026-03-29T10:52Z	⚠️ Scope note

Missing Reports (present 2026-03-27, absent today and 2026-03-28):

• ❌ Observability Coverage Report
• ❌ Daily Performance Summary
• ❌ Daily Code Metrics Report
• ❌ DeepReport Intelligence Briefing
• ❌ Prompt Clustering Analysis

🔍 Data Consistency Analysis

Cross-Report Metrics Comparison

Metric	Team Evolution	Copilot Agent Analysis	Scope Match	Status
Total PRs (24h)	20 (all contributors)	29 (agent_prs_total)	⚠️ Different scopes	ℹ️ See Note 1
Merged PRs (24h)	16 (all contributors)	23 (agent_prs_merged)	⚠️ Different scopes	ℹ️ See Note 1
Merge rate (24h)	80%	79.3%	✅ Consistent	✅
Firewall runs (7d)	—	—	—	—
Safe output items	66 processed	—	N/A	✅

Scope Notes:

• Note 1 (PR counts): Team Evolution reports 20 total / 16 merged PRs for a general "last 24h all-contributor" view, while Copilot Agent Analysis shows 29 total / 23 merged for the specific agent-PRs window (2026-03-28 11:30Z – 2026-03-29 11:30Z). The Team Evolution report describes most activity as AI-authored, yet shows ~7 fewer total PRs. This may reflect different time window boundaries or Team Evolution filtering to "notable" PRs. The merge rates are consistent (80% vs 79.3%), suggesting the ratio is reliable even if totals differ. Scope clarification recommended.
• Note 2 (workflow run counts): Safe Output Health analyzed 42 runs (last 24h all workflows), while Firewall analyzed 51 total runs (7-day window, firewall-enabled only). These are different scopes and not comparable.
• Note 3 (DIFC scope): DIFC report covers a 3-day analysis window (2026-03-27 to 2026-03-29) — not strictly "last 24h". The total 221 events span this window.

Consistency Score

• Overall Consistency: 85% (metrics with same scope are aligned)
• Critical Discrepancies: 0
• Minor Discrepancies: 1 (PR total count — scope difference)
• Scope mismatches documented: 3

⚠️ Issues and Anomalies

Critical Issues

None identified.

Warnings

1. 21 Workflow Compile Failures (Static Analysis)

   • Affected Report: Static Analysis #23478
   • Metric: workflows_compiled out of workflows_scanned
   • Description: 21 of 178 scanned workflows (11.8%) failed to compile, all due to the same missing shared import file
   • Expected: 0 compile failures (all workflows should compile)
   • Actual: 21 failures — systemic, single root cause
   • Severity: High
   • Recommended Action: Identify and restore the missing shared import file; add a CI check to prevent import regressions

2. High Run Failure Rate in Safe Output Health

   • Affected Report: Safe Output Health #23459
   • Metric: Run-level success rate vs item-level success rate
   • Description: 16 of 39 completed runs failed (41% run failure rate), though per-item safe output success rate is ~94% (62/66 items). The discrepancy between run-level and item-level success suggests run failures don't necessarily prevent safe output execution.
   • Severity: Medium
   • Recommended Action: Distinguish between run failures caused by the agent task (expected) vs infrastructure/configuration failures

3. PR Count Scope Discrepancy (Team Evolution vs Agent Analysis)

   • Affected Reports: Team Evolution #23427, Copilot Agent Analysis #23438
   • Metric: Total PRs, merged PRs in 24h window
   • Description: Team Evolution reports 20 total / 16 merged while Agent Analysis reports 29 total / 23 merged. Merge rates are consistent (~80%) but absolute counts differ.
   • Severity: Low
   • Recommended Action: Align time window boundaries and document scopes explicitly in both reports

4. 5 Missing Report Types

   • Affected: Observability Coverage, Daily Performance Summary, Daily Code Metrics, DeepReport, Prompt Clustering
   • Description: These reports were present on 2026-03-27 but absent on both 2026-03-28 and 2026-03-29, suggesting scheduled workflows may have stopped or failed
   • Severity: Medium
   • Recommended Action: Verify that scheduled workflows for these reports are still enabled and running

Data Quality Notes

• DIFC report covers 3-day window (not strictly 24h) — labels as daily but analyzes longer window
• Team Evolution uses qualitative language ("~25 commits") rather than precise counts — reduces comparability
• Copilot Session Insights 20% completion rate is improving from March lows (0–14%), no anomaly

📈 Trend Analysis

Day-over-Day Comparison (Key Operational Metrics)

Metric	2026-03-29	2026-03-28	Change
Agent PRs total	29	30	-3.3%
Agent PRs merged	23	24	-4.2%
Agent merge rate	79.3%	80.0%	-0.9%
Agent avg duration	171 min	191 min	-10.5% ↑
Firewall block rate	1.3%	N/A (7d window)	—
Session completion rate	20%	25%	-5pp
Session action_required	48%	—	elevated

Notable Trends

• Agent throughput is stable: ~29-30 PRs/day with ~80% merge rate maintained over 3 days
• Resolution speed improving: Average merge time dropped from 629 min (2026-03-27) to 171 min (2026-03-29), a 73% improvement
• Session completion recovering: From 0–14% mid-March to 20–25% late March — positive trajectory
• Security activity spike: DIFC integrity events rising (51→92→78 over 3 days), driven by Auto-Triage Issues workflow processing community submissions

📝 Per-Report Analysis

Sergo Error Patterns (#23480)

Time Period: Today (2026-03-29) — baseline run for this repository
Quality: ✅ Valid

Metric	Value	Validation
Go files analyzed	629 non-test	✅
LOC analyzed	~149K	✅
Total findings	3 (1 High, 1 Med, 1 Low)	✅
Tasks generated	3	✅

Notes: First Sergo run for this repo — establishes baseline. High-priority finding (timestamps in error strings) affects 35+ call sites.

DIFC Integrity-Filtered Events (#23479)

Time Period: 2026-03-27 to 2026-03-29 (3-day window)
Quality: ✅ Valid — Note: labeled "daily" but covers 3 days

Metric	Value	Validation
Total filtered events	221	✅
Unique users blocked	26	✅
Unique workflows affected	10	✅
Most active workflow	Auto-Triage Issues (60%)	✅

Static Analysis (#23478)

Time Period: 2026-03-29 scan
Quality: ⚠️ Issues — 21 compile failures

Metric	Value	Validation
Workflows scanned	178	✅
Compiled successfully	157 (88.2%)	⚠️ 21 failures
zizmor issues	3,455	✅
actionlint issues	3,928	✅
Template injection (High)	24	⚠️ Needs attention

Safe Output Health (#23459)

Time Period: Last 24 hours (2026-03-29)
Quality: ⚠️ Notable run failure rate

Metric	Value	Validation
Runs analyzed	42	✅
Completed runs	39	✅
Successful runs	23 (59% run rate)	⚠️ Low run success rate
Safe output items	66 total	✅
Item success rate	~94% (62/66)	✅ Healthy
Failure clusters	3 identified	✅

Daily Firewall (#23441)

Time Period: 7-day scope through 2026-03-29
Quality: ✅ Valid

Metric	Value	Validation
workflow_runs_analyzed	51 runs (44 unique workflows)	✅
firewall_requests_total	1,206	✅
firewall_requests_allowed	1,190 (98.7%)	✅
firewall_requests_blocked	16 (1.3%)	✅
firewall_domains_blocked	6	✅

Copilot Agent Analysis (#23438)

Time Period: 2026-03-28 11:30Z – 2026-03-29 11:30Z
Quality: ✅ Valid

Metric	Value	Validation
agent_prs_total	29	✅
agent_prs_merged	23	✅
agent_success_rate	79.3%	✅ Math check: 23/29 = 79.3% ✅
Avg duration	171 min	✅

Internal consistency: Math verified — 23 merged + 5 closed + 1 open = 29 total ✅

Copilot Session Insights (#23444)

Time Period: 2026-03-29 (burst window 11:15–11:22 UTC)
Quality: ✅ Valid

Metric	Value	Validation
Sessions analyzed	50	✅
Successful	10 (20%)	✅
Action required	24 (48%)	⚠️ Elevated but explained
Failed/abandoned	2 (4%)	✅
Skipped + in-progress	14 (28%)	✅

Internal consistency: 10 + 2 + 13 + 1 + 24 = 50 ✅

Team Evolution Insights (#23427)

Time Period: Last 24 hours (Mar 28–29)
Quality: ⚠️ Scope note (approximate counts, different PR scope)

Notes: Uses approximate figures (~25 commits). PR counts (20 total, 16 merged) differ from Copilot Agent Analysis totals — likely reflects different population (all contributors vs agent-only) and/or time window. Merge rates align (80% vs 79.3%).

💡 Recommendations

Process Improvements

1. Standardize PR count scopes: Both Team Evolution and Copilot Agent Analysis should explicitly document their PR population (all contributors vs agent-only) and use consistent time window boundaries to enable direct comparison
2. Use precise counts in Team Evolution: Replace approximate figures (~25 commits) with exact counts to enable regulatory comparison
3. Align DIFC reporting window: Report should clarify it covers 3 days, not 1 day, to avoid confusion

Data Quality Actions

1. Restore missing workflows: Investigate why 5 report types (Observability, Performance, Code Metrics, DeepReport, Prompt Clustering) have been absent for 2 days — check scheduled workflow status
2. Fix shared import regression: The 21 workflow compile failures in Static Analysis all trace to the same missing import — fix should be straightforward and has high impact

Workflow Suggestions

1. Add compile health gate: CI should fail if workflow compile failure rate exceeds 5% to catch import regressions early
2. Separate run vs item metrics: Safe Output Health should more prominently distinguish run-level failure rate (41%) from item-level success rate (94%) to avoid misinterpretation

📊 Regulatory Metrics

Metric	Value
Reports Reviewed	8
Reports Passed (✅)	5
Reports with Issues (⚠️)	3
Reports Failed (❌)	0
Overall Health Score	81%
Critical Discrepancies	0
Warnings Issued	4
Previous Reports Closed	2 (#23275, #23172)

---

Report generated automatically by the Daily Regulatory workflow
Workflow run: §23719016196
Data sources: Daily report discussions from github/gh-aw (2026-03-29)
Metric definitions: scratchpad/metrics-glossary.md

AI generated by Daily Regulatory Report Generator · history

• expires on Apr 1, 2026, 9:10 PM UTC

[github-actions[bot]]

This report has been superseded by a newer daily regulatory report for 2026-03-30.

[github-actions[bot]]

This report has been superseded by a newer daily regulatory report.