[detection-analysis] Detection Analysis Report — 2026-07-13 (last 24h) #45269
Closed
Replies: 1 comment
-
|
This discussion has been marked as outdated by Detection Analysis Report. A newer discussion is available at Discussion #45488. |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
Summary
Warning
5 workflow(s) flagged with detection misconfigurations. See the section below.
Read: detection-enabled and regular runs land at an identical 82.9% success rate this window, so detection adds no measurable reliability penalty. Detection runs cost ~1.6× the tokens of regular runs on average (137,778 vs 85,436), reflecting the extra detection-pass overhead.
Comparison Chart
Misconfigured Workflows
All 5 flagged workflows are audit/analysis-type workflows running without
gh-aw-detection: true— the workflow name matches an audit/analysis keyword but detection is absent from itsfeatures.No workflows showed inconsistent (alternating) detection state, and no active workflow (>3 runs) had detection explicitly disabled —
Smoke CIruns withgh-aw-detection: falsebut sits at exactly 3 runs, just below the flag threshold, so it is surfaced in Recommendations rather than flagged.View All Run Metrics (70 workflows)
Detection column: Enabled =
gh-aw-detection: true; Disabled = explicitlyfalse; Absent = feature not set; Mixed = both within the window. Avg tokens = input+output;0means no usage artifact was recorded for that run.View Historical Trend (17 days)
Detection-run volume has trended upward over the window (from single digits in late June to 35 runs today), while detection success rate has stayed in the 80–95% band apart from a 2026-07-09 dip.
Recommendations
gh-aw-detection: trueto each workflow's frontmatter — these are exactly the workflows where detection provides the most value.Smoke CI(3 runs,gh-aw-detection: false, 0% success). It is one run below the active-workflow threshold and is currently failing every run; confirm whether detection should be enabled and why it is failing.Changeset Generator,Claude Code User Documentation Review,Copilot PR Conversation NLP Analysis,daily-experiment-report). This report used only usage artifacts; re-run with--artifacts agentto confirm whether any failure originated in the detection job step itself (misconfiguration rule 3) rather than the agent body.Caveats: Misconfiguration rule 1 (detection disabled on workflows with >3 runs over 7 days) was evaluated against the 24h window as a proxy — a full 7-day pull would refine it. Rule 3 (detection-job failures) could not be confirmed because only usage artifacts were downloaded.
References:
Warning
Firewall blocked 1 domain
The following domain was blocked by the firewall during workflow execution:
awmgmcpgSee Network Configuration for more information.
Beta Was this translation helpful? Give feedback.
All reactions