A demo GitHub Action for testing the Origin supply-chain-security platform.
- Logs
Hello, <name>!(default name:world). - Unconditionally writes credential-shaped strings to the job log so the Origin credential scanner has something to detect.
The credential-leak behavior is tied to the action SHA: pre-leak SHAs run cleanly; this SHA and any newer ones emit the leak. This lets a demo distinguish "clean SHA → job allowed" from "leaky SHA → post-job scanner findings → job blocked" without runtime configuration.
Reference an earlier SHA (before the credential-leak step was added) to demo the happy path:
jobs:
demo:
runs-on: self-hosted-with-origin
steps:
- uses: resilientscale/gh-demo-action@<pre-leak-sha>
with:
name: 'GRunS'Reference the current SHA to demo the credential-scanner path:
jobs:
demo-leak:
runs-on: self-hosted-with-origin
steps:
- uses: resilientscale/gh-demo-action@<leak-sha>The leak step emits randomly-generated values matching common secret patterns: GitHub PAT (ghp_…), AWS access key ID (AKIA…), AWS secret access key, Slack bot token (xoxb-…), Stripe live key (sk_live_…). They are not real credentials.
| Name | Required | Default | Description |
|---|---|---|---|
name |
no | world |
Name to greet in the hello-world step |
This action is intended for demo and testing of Origin only.
- The credential-leak step deliberately writes secret-shaped strings to logs.
- Do NOT use any SHA of this action in production workflows.
- The repository should NOT be granted access to real secrets.
See gh-runner-sentinel/setup/demo-thread.md — the three scenarios this action exercises:
- Happy path — workflow uses a clean (pre-leak) SHA of this action; agent reports the action; GRunS allows it; job runs to completion.
- Immutable-action enforcement — workflow references this action by tag (e.g.
@v1) instead of a SHA, withGRUNS_REQUIRE_IMMUTABLE_ACTIONS=trueon the agent; agent submits the mutable ref and GRunS blocks the job with reason "mutable action reference". - Post-run credential audit — workflow uses the leak SHA of this action with
GRUNS_SCAN_LOGS_FOR_CREDENTIALS=true+GRUNS_FAIL_JOB_ON_LEAK=trueon the agent; agent runs the post-job credential scanner, finds the synthesized secrets, reports findings, fails the job.