ENGINEERING
HuggingFace card-drift watcher to Linear eval task
Polls the HuggingFace cards of the open models your team uses on a schedule, and when a model card or license changes it opens a Linear eval task pre-filled with the diff.
How it runs
The automated pipeline, trigger to output.
- TriggerDaily schedule fires
- ActionFetch HuggingFace card + metadata for each watched modelHugging Face
- ActionLoad last snapshot from PostgresPostgres
- LogicDiff card/license/tag fields; skip if unchanged
- OutputOpen Linear eval task with the diffLinear
- ActionWrite new snapshot to PostgresPostgres
What it does
Keeps a watchlist of the open-weight models your product depends on (for example `meta-llama/Llama-3.1-8B-Instruct`) and checks each one's HuggingFace model card on a fixed cadence. When the card body, license, or pipeline tag changes versus the last snapshot, it files a Linear issue describing exactly what moved so an engineer can re-run evals before the change reaches production.
When to use it
Use it when you ship features on top of third-party open models and a quiet upstream edit (a relicense, a new usage restriction, a changed recommended prompt format) could silently break or legally compromise your stack. It turns "someone noticed on Twitter" into a tracked task.
How it works
- 1A daily schedule fires the run.
- 2For each watched model it fetches the current HuggingFace card metadata and README.
- 3It compares the new content against the stored snapshot to compute a field-level diff.
- 4A branch checks whether anything meaningful changed; unchanged models are skipped.
- 5For changed models it creates a Linear issue titled with the model id and the changed fields, embedding the before/after diff.
- 6It writes the new snapshot back so the next run compares cleanly.
Set it up
What you configure once, before turning it on.
- 1Connect Hugging FaceModels, datasets, spaces — the open-source hub.
- 2Connect LinearIssues, projects, cycles, triage.
- 3Connect PostgresAny Postgres URL — query, write, migrate.
- 4Set each agent's modelWe leave models unset so you pick the tier — fast + cheap, or top-quality.
- 5Tune it to your dataEdit the prompts, filters, and field mappings so it matches how your team works.
- 6Test, then turn it onRun once against a sample, confirm the output, then enable the trigger.
More Engineering workflows
Gate breaking API PRs behind downstream consumer acknowledgement
When a PR introduces a breaking contract change, comments the impact summary back on the PR, applies a blocking label.
Publish a versioned API changelog to Confluence on each release tag
On a new semver release tag, gathers the contract changes since the last release and writes a clean.
Agent reviews model-license fit and suggests compliant swaps on the PR
When a PR adds a Hugging Face model, an agent reads the model card and license, judges fit against your commercial-use policy.
Upgrade Impact Router to Module Code Owners
Maps a dependency-bump PR's affected modules to their CODEOWNERS, then DMs each owner on Slack with only the changelog slice that touches code they own.
Re-Voice IVR Prompts on Phone-Tree Config Merge
When a phone-tree config change merges in GitHub, regenerates the ElevenLabs audio for any prompt whose script changed in the diff and opens a follow-up PR adding the new audio…
Upstream Release to Notion Upgrade Brief
When a watched package publishes a new release, fetches the release notes, maps them to the internal modules that depend on it.
Run it inside a business
This workflow drops into a full company template. Import the org, and this is one of the playbooks its agents run.
Run this workflow in your colony.
14-day trial. No DevOps. No Sales call. Provisioned in under a minute.