Skip to content

Enable gh-aw-detection on 50% of agentic workflows #40698

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
pelikhan merged 1 commit into from
Jun 21, 2026
Merged

Enable gh-aw-detection on 50% of agentic workflows #40698

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
84 changes: 26 additions & 58 deletions .github/workflows/agent-performance-analyzer.lock.yml
View file
Open in desktop

Large diffs are not rendered by default.

3 changes: 2 additions & 1 deletion .github/workflows/agent-performance-analyzer.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -61,7 +61,8 @@ experiments:
notify:
issue: 33280
issue: 33280

features:
gh-aw-detection: true

@github-actions github-actions Bot Jun 21, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

[/zoom-out] The 50% rollout selects workflows alphabetically (A–D range visible in this diff). Alphabetical order is not random, so the cohort may share structural characteristics — e.g. all workflows whose names start with letters near the top of the alphabet could cluster by creation era, team, or trigger pattern.

A few questions worth confirming before merging:

  • Is there monitoring/alerting that will surface detection failures specifically for the new binary path vs. the inline-engine path?
  • Is there a rollback plan documented if the 50% cohort starts producing systematic false positives or detection errors?
  • If results look good, what is the intended next step (100% rollout, or graduated increases)?
💡 On sampling strategy

For a statistically representative sample, random shuffling before slicing at 50% is preferable to alphabetical selection. Alphabetical order can introduce bias if there is any correlation between workflow name and workflow characteristics (e.g., daily-* workflows are all scheduled and may behave differently from event-driven workflows). That said, alphabetical is operationally convenient and fine for a first canary pass — just worth noting if the comparison results are used to make conclusions about parity.

---

{{#runtime-import? .github/shared-instructions.md}}
Expand Down
67 changes: 24 additions & 43 deletions .github/workflows/agent-persona-explorer.lock.yml
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

2 changes: 2 additions & 0 deletions .github/workflows/agent-persona-explorer.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -54,6 +54,8 @@ imports:


- shared/otlp.md
features:
gh-aw-detection: true
---

# Agent Persona Explorer
Expand Down
2 changes: 1 addition & 1 deletion .github/workflows/ai-moderator.lock.yml
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

4 changes: 2 additions & 2 deletions .github/workflows/ai-moderator.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -57,8 +57,8 @@ safe-outputs:
allowed-reasons: [spam]
threat-detection: false
checkout: false


features:
gh-aw-detection: true
---

# AI Moderator
Expand Down
Loading
Loading