[archivx] Agentic Workflow Visual Summary — 2026-08-24

[github-actions[bot]]

Overview

Over the analyzed window (2026-08-23 14:39 UTC → 2026-08-24 02:09 UTC, 203 runs with complete metadata), the agentic workflow fleet held an 88.2% success rate (179 succeeded, 24 failed). Activity was dominated by PR-facing automation — PR Sous Chef alone accounted for 33 runs with zero failures — while review/skill-check workflows generated the bulk of remaining volume. Two workflows stand out for reliability concerns: Code Scanning Fixer failed both of its 2 runs, and Design Decision Gate 🏗️ failed 3 of 12 runs (25%).

---

Key Metrics

Metric	Value
Total Runs	203
Success Rate	88.2%
Top Engine	copilot (90 runs, 44%)
Most Active Workflow	PR Sous Chef (33 runs)
Avg Tokens / Run	~44,477

---

Animated Diagram

The animated architecture diagram is attached as workflow artifact archivx-animated-diagram.

The HTML file opens directly in any browser. It includes a ☀/☾ light-dark toggle and a ⏯ pause button, and honors reduced-motion preferences.

---

Top Failures

View Details

Workflow	Failures	Total Runs	Failure Rate
Design Decision Gate 🏗️	3	12	25%
Code Scanning Fixer	2	2	100%
Issue Arborist	1	2	50%

Code Scanning Fixer's 100% failure rate (on low volume — just 2 runs) is worth a closer look before it scales up; Design Decision Gate has the most absolute failures and the most run volume among the three.

---

Workflow Activity

View Details

Workflow	Runs	Success Rate
PR Sous Chef	33	100%
Impeccable Skills Reviewer	14	93%
Test Quality Sentinel	13	92%
Matt Pocock Skills Reviewer	13	92%
Ponytail Reviewer	13	92%
PR Code Quality Reviewer	13	92%
Design Decision Gate 🏗️	12	75%
PR Description Updater	10	100%
Issue Monster	7	100%
Auto-Triage Issues	6	100%

Engine distribution across all 203 runs: copilot 90, pi 59, claude 26, codex 22, aider 2, crush 1, goose 1, opencode 1, gemini 1.

Generated by 🌅 Archivx — Animated Workflow Visualizer · claude · agent · 80.4 AIC · ⌖ 9.32 AIC · ⊞ 7.3K · ◷

• expires on Sep 22, 2026, 6:26 PM UTC-08:00