atlas-ci-release
Skill2 starsChanged 5 months agoApache-2.0View on GitHub
Manage CI bundle releases: manual dispatch, tag naming, monitoring the build-bundle workflow, promoting a release, and rolling back. Use when releasing a new bundle, debugging a CI failure, or rolling back to a previous bundle.
- Commits and pushes
In this repository16 other files it gives its agents.
- architecture-decision-recordsSkill
- atlas-add-toolSkill
- atlas-build-bundleSkill
- atlas-optimization-patternsSkill
- atlas-rag-evaluationSkill
- embedding-strategiesSkill
- hybrid-search-implementationSkill
- mcp-builderSkill
- python-code-styleSkill
- python-observabilitySkill
- python-resilienceSkill
- python-testing-patternsSkill
- python-type-safetySkill
- readme-maintenanceSkill
- skill-creatorSkill
- uv-package-managerSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 4 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 35
Do NOT dispatch from main if there are unmerged changes that affect the bundle content.
- Distinctive: line 64
If the workflow exceeds 60 minutes, check the runner logs for:
- Distinctive: line 71
When the build workflow succeeds:
- Distinctive: line 95
No code rollback is needed — bundles are independent artifacts.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
When to use
- User asks to "release a new bundle", "trigger a CI build", "create a release"
- User wants to "tag a release", "cut a release"
- A CI build failed and needs investigation
- User wants to "roll back to the previous bundle"
Not for
- Building the bundle locally (use
atlas-build-bundle) - Changing the CI workflow itself (edit
.github/workflows/build-bundle.ymlmanually)
Workflow
1. Manual dispatch
Go to GitHub → Actions → Build Bundle → Run workflow.
Parameters:
| Parameter | Value | When |
model_id | Xenova/bge-small-en-v1.5 | Default — production builds |
ref | main | Stable releases |
ref | <branch> | Testing a feature branch |
Do NOT dispatch from main if there are unmerged changes that affect the bundle content.
2. Tag naming convention
Tags are auto-generated from the workflow run: bundle-<YYYYMMDD>-<run_number> (e.g., bundle-20260622-42).
For manual tags (when the auto-tag is insufficient):
git tag bundle-<YYYYMMDD>-<run_number>-<reason>
git push origin bundle-<YYYYMMDD>-<run_number>-<reason>Reasons: hotfix, rollback, rc1, rc2.
3. Monitor the workflow
The workflow takes ~30–60 minutes.
| Phase | Duration | Key log lines |
| Setup | 1 min | Setup uv, Cache npm |
| Clone docs | 2–3 min | Cloning servicenow-docs |
| Chunk | 2–5 min | Chunked N files |
| Download model | 1–3 min | Downloading Xenova/bge-small-en-v1.5 |
| Embed | 15–30 min | Embedding chunk N/M |
| Package | 1 min | Creating tarball |
| Smoke test | 2 min | Smoke test passed |
| Publish | 1 min | Uploading release asset |
If the workflow exceeds 60 minutes, check the runner logs for:
- Slow model download (HF mirror issues)
- OOM (runner has 2 vCPUs / 7 GB)
- Stuck chunking (infinite loop in a single file)
4. Promote a release
When the build workflow succeeds:
- The tarball is uploaded as a GitHub Release asset.
- The release name matches the tag.
- To make it the "latest" release:
gh release edit <tag> --latest
5. Roll back
To revert to a previous bundle:
# Find the previous release's tag
gh release list --limit 10
# Download the previous bundle
gh release download <previous-tag> --pattern '*.tar.zst'
# Make it the latest
gh release edit <previous-tag> --latestNo code rollback is needed — bundles are independent artifacts. Just re-point consumers.
CI workflow reference
File: .github/workflows/build-bundle.yml
| Key detail | Value |
| Trigger | Monthly cron (0 6 1 * *) + workflow_dispatch |
| Runner | ubuntu-latest (2 vCPU, 7 GB, no GPU) |
| Model | Xenova/bge-small-en-v1.5 (int8, 384-dim) |
| Backend | ONNX+CPU (forced via --backend onnx-cpu) |
| Batch size | 256 |
| Env | HF_TOKEN set from secrets |
| Post-build | atlas-smoke |
| Artifact | GitHub Release (auto-tag) |
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

