[archivx] Agentic Workflow Visual Summary — 2026-10-19

[github-actions[bot]]

Overview

Across the analyzed runs, agentic workflows in this repository are healthy: a 94.1% success rate with copilot as the dominant engine and PR Sous Chef as the single most active workflow. Coverage caveat: the mission asked for a 7-day window, but log collection hit a GitHub API rate limit after the initial fetch and one continuation attempt, so the analyzed data actually spans only about 10 hours (2026-10-18 22:33 UTC – 2026-10-19 08:39 UTC), not the full week. Treat these numbers as a recent snapshot, not a 7-day trend.

---

Key Metrics

Metric	Value
Total Runs	290 (with metadata, of 389 run directories collected)
Success Rate	94.1%
Top Engine	copilot (124 runs)
Most Active Workflow	PR Sous Chef (23 runs)

---

Diagram

The SVG architecture diagram has been committed to docs/src/assets/archivx/agentic-workflows-archivx.svg in the generated pull request.

The SVG can be embedded directly in docs pages using standard markdown image syntax.

---

Top Failures

View Details

Failures were spread thin — 17 total across 290 analyzed runs, mostly single-occurrence smoke/engine-validation checks rather than a repeating systemic issue:

Workflow	Failures / Runs
Contribution Check	1 / 2
Daily Storify	1 / 1
Daily Container Image Security Scan	1 / 1

The remaining 14 failures were each isolated to a single run of a distinct "Smoke *" engine-harness workflow (e.g. Smoke Gemini, Smoke DeepSeek Harness, Smoke Copilot MAI, Smoke Drive, Smoke Azure PI, Smoke Crush, Smoke Goose) — these exist specifically to validate individual engine/model routes, so isolated failures there are plausibly flaky infra checks rather than product regressions. No cluster had enough repeated occurrences in the collected window to point at a specific root cause with confidence.

---

Workflow Activity

View Details

Workflow	Runs	Success Rate
PR Sous Chef	23	87%
Test Quality Sentinel	18	100%
Ponytail Reviewer	18	100%
Issue Monster	18	100%
Matt Pocock Skills Reviewer	17	100%
PR Code Quality Reviewer	17	100%
Design Decision Gate 🏗️	16	100%
Avenger	9	100%
Daily Go Test Parallelizer	5	100%
Smoke Copilot Routed	3	100%

---

Coverage & validation notes

• Data collection: 389 run directories fetched (290 with usable aw_info.json/run_summary.json), limited by a GitHub API rate limit on the second continuation call — full 7-day coverage was not achieved.
• Diagram geometry: check_diagram.py passed with 0 violations after adding explicit per-category tiers to the architecture graph.
• Diagram readability (contrast/clipping in-browser): not verified — every installed browser channel (chrome, msedge, bundled chromium) fails to launch in this sandbox due to a GLIBC_2.43/libasound.so.2 mismatch under /run/awf-lib, which is sandbox infrastructure outside this workflow's control. This is disclosed in the pull request body as well.

Generated by 🌅 Archivx — Workflow Visualizer · claude · agent · 185.2 AIC · ⌖ 1.12 AIC · ⊞ 6.7K · ◷

• expires on Nov 18, 2026, 1:00 AM UTC-08:00