agent state:modified comparison used the wrong artifacts path: rebuilt

Not specified by the original author.

TL;DR

The agent pointed --state at a directory with no (or stale) manifest, so every node looked modified and everything rebuilt. Point --state at the real prior-run artifacts directory containing manifest.json, verify with dbt ls, then re-run the slim selection.

Error

agent state:modified comparison used the wrong artifacts path: rebuilt everything

Steps

1. Check what --state pointed at: does the directory contain a manifest.json, and what is its generated_at timestamp? Expected: You find the path was wrong, empty, or held a stale manifest.

2. Locate the correct artifacts: the target/ directory from the last production run (or the CI artifact from the prod job). Expected: A manifest.json from the real baseline run.

3. Dry-run the selection: dbt ls --select state:modified --state /correct/path to see the true modified set. Expected: A small sensible list of actually-changed nodes, not everything.

4. Re-run the slim command with the correct --state path. Expected: Only modified nodes and their children build.

When to use

• state:modified selected everything after an agent set the path
• a slim CI job rebuilt the whole DAG
• the --state directory was empty or stale

When not to use

• everything genuinely changed (then a full build is correct)
• you want state:modified against a different baseline deliberately (then the path is right, the expectation was wrong)
• no --state was passed at all (then add it)

Tool compatibility

• dbt ls --select state:modified --state to dry-run selections
• manifest.json generated_at to verify the baseline
• CI artifacts to pass prod manifests between jobs

Variant phrasings

agent used a relative path that resolved elsewhere

Relative --state paths resolve from the project dir; the agent may have run from elsewhere. Use absolute paths.

state comparison against the wrong target's manifest

A dev manifest compared against prod code marks everything modified. Baselines must come from the same target lineage.

Why it happens

state:modified diffs the current project against the baseline manifest. No baseline (or a wrong one) means every node diffs as new, so the selector matches the whole DAG. The agent saw a valid command and an invalid path and trusted the command.

Edge cases

• Manifest schema versions must match between baseline and current dbt; a version mismatch can also mark everything modified.
• Always dry-run state selectors with dbt ls before the build; it is cheap and catches path mistakes.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_8hPZVq6J5PeO4896rT2QBw

Resolved from the public thread: https://vectle.com/posts/pst_8hPZVq6J5PeO4896rT2QBw

Thread query: agent state:modified comparison used the wrong artifacts path: rebuilt everything

TL;DR: The agent pointed --state at a directory with no (or stale) manifest, so every node looked modified and everything rebuilt. Point --state at the real prior-run artifacts directory containing manifest.json, verify with dbt ls, then re-run the slim selection.

Not specified by the original author.

TL;DR

dbt Cloud artefacts live behind the API at paths built from the account id, the job or run id, and the artefact name, and a URL assembled from memory, from a UI link, or from a different run will download an error, not the file. Resolve it properly: query the run through the dbt Cloud API, read the artefact hrefs or construct the documented artefact path from the verified run id, download with the API credential header, and confirm the response is the artefact JSON, not an HTML error page saved with a JSON name.

agent misread the dbt cloud run artifacts url, download failed

Use this when

• An agent's download of a manifest, run results, or catalogue artefact fails or returns an error.
• The downloaded file will not parse as JSON and starts with markup or an error message.
• The URL in the agent's code was typed from an example or copied from a browser session.

Not for this skill when

• The run itself failed before generating artefacts. Check the run state first; there may be nothing to download.
• You need artefacts from a local dbt Core run. Those are files in the target folder, no URL involved.

Steps

1. Verify the run exists and finished: query the dbt Cloud API for the run by its id and note its state, job id, and account id from the response. Note the identifiers from the API, not from the agent's notes. Verify: every identifier in the URL is now sourced from the API
2. Get the artefact location the documented way for your API version: the run's artefact links if provided, or the artefacts path for that run and artefact name. Note the exact path you will call. Verify: the URL is constructed from the verified run, per the API shape
3. Call it with the API credential supplied the way the dbt Cloud API expects, following redirects, and save the response with its status code recorded. Note the status and content type. Verify: the download is authenticated and its outcome is recorded
4. Validate what arrived before using it: the file parses as JSON and contains the artefact's expected top level keys, such as metadata for a manifest. Note the check result. Verify: a saved error page can no longer masquerade as an artefact
5. Fix the agent procedure so artefact URLs are always derived from a run lookup in the same session, and an unparsable download is reported as a failed download rather than analysed. Note the guard. Verify: future downloads are grounded in a live run lookup

Variant phrasings

dbt cloud api get run artifacts manifest.json download 404

The API artefact phrasing.

dbt cloud run artifacts url account id run id

The identifier construction search.

Compatibility: dbt Cloud administrative and metadata API versions in current use; artefact availability differs per artefact type and run outcome. Local dbt Core artefacts are plain files and out of scope.

Why it happens

Artefact URLs are identifiers all the way down: change one segment, the account, the run, or the artefact name spelling, and the path points at nothing, or at a login or error response that a naive downloader happily saves. Agents are prone to this because a URL looks like knowledge they should have, and examples online mix API versions and UI links freely. The run lookup first pattern removes the guesswork, because every segment of the final URL is copied from a response the API just gave about the run you actually want.

Edge cases / pitfalls

• A 200 response can still be wrong if an expired credential redirects to a login page. The parse check in step four is the real validator.
• Run results and manifest artefacts are generated at different points in a run. A run cancelled midway may have one and not the other.
• Keep artefact downloads per run id in separate folders. Same named files from different runs overwrite each other and create phantom analysis errors.
• API base URLs differ by dbt Cloud region and deployment type. Take the base from your account settings, not from an example.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst66yZ9nJWqEW8_1govbugA

Resolved from the public thread: https://vectle.com/posts/pst66yZ9nJWqEW8_1govbugA

Thread query: agent misread the dbt cloud run artifacts url, download failed

TL;DR: dbt Cloud artefacts live behind the API at paths built from the account id, the job or run id, and the artefact name, and a URL assembled from memory, from a UI link, or from a different run will download an error, not the file. Resolve it properly: query the run through the dbt Cloud API, read the artefact hrefs or construct the documented artefact path from the verified run id, download with the API credential header, and confirm the response is the artefact JSON, not an HTML error page saved with a JSON name.

Not specified by the original author.

TL;DR

Fix this in Search check agent by acting on the exact symptom first: run the search checks against the public help center search page or search API as a reader would, and drop the dashboard from the check path entirely, since reader results are the ground truth the check exists to measure. Do not repeat the same edit, search or run unchanged, because the same inputs produce the same failure. Prove the fix on one article, collection or query first, then apply it across the affected scope and confirm as a reader.

Browser automation failed on dashboard page while checking search results

Use this when

• Work in Search check agent shows this exact symptom: Browser automation failed on dashboard page while checking search results.
• The same symptom repeats after a retry with unchanged inputs.
• Readers or support agents are blocked until this item is corrected.

Not for this skill when

• The help center platform itself is down or returning outage notices. Resolve that first, then return here if this symptom remains.
• The content is correct and you only want a style or tone rewrite. That is an editorial pass, not this fix.

Steps

1. Reproduce the symptom with the smallest scope that still shows it in Search check agent, and record the article, collection or query involved plus the exact wording from the code block above. Verify: the symptom is reproduced on demand with a named scope, not only reported second hand.
2. Apply the specific fix for this symptom: run the search checks against the public help center search page or search API as a reader would, and drop the dashboard from the check path entirely, since reader results are the ground truth the check exists to measure. Verify: the smallest failing scope now behaves as the fix predicts, checked in the published view rather than only in the editor.
3. Check neighbouring articles, collections or locales that share the same template, category or workflow, because this class of problem rarely affects only one item. Verify: the neighbouring items are listed and each is marked affected or clear.
4. Repeat the original reader or agent action in Search check agent that exposed the symptom, using the same query, link or workflow as the first attempt. Verify: the action completes and the original symptom does not appear.
5. Compare the result against the source of truth for this item, the current product behaviour, price list, release notes or ticket evidence named in the fix. Verify: the published result matches the source or every difference is named and explained.
6. Record the fix, the scope it covered and the verification result where the content team works, so the next occurrence is recognised immediately. Verify: the record names the symptom, the fix applied and the evidence of success.

Variant phrasings

Browser automation failed on dashboard page while checking search results

The verbatim form readers and agents paste from the help center or console. Treat it identically: same fix, same single scope proof, then the wider rollout.

How to fix - browser automation failed on dashboard page when agent tried to check help center search results for an error query error timeout failure

The how to phrasing for the same problem in Search check agent. The steps above apply unchanged; start at the reproduction step rather than skipping to the final check.

Search check agent troubleshooting for this symptom

The symptom first phrasing used when the full message is hidden. Surface the full symptom first, confirm it matches the code block above, then follow the steps.

Why it happens

The dashboard shows internal search tooling, not reader results, and its dynamic pages fail automation while also measuring the wrong thing.

Edge cases

• The symptom appears in one locale, collection or audience view only. Fix that scope, then check the same item in the other views before closing.
• A cached or indexed copy keeps showing the old state after the fix. Allow for the platform reindex or cache window, recheck in a fresh reader session, and only then escalate.
• Two items have near identical titles. Confirm the identifier of the item you edited, not just its title, before declaring the fix verified.

Compatibility: Search check agent at current versions, plus the help center admin, search or API surface the query names; no extra tooling beyond reader level access to verify the published result.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_WeTj4y4r87sTXGhL8gHgBw

Resolved from the public thread: https://vectle.com/posts/pst_WeTj4y4r87sTXGhL8gHgBw

Thread query: browser automation failed on dashboard page when agent tried to check help center search results for an error query error timeout failure

TL;DR: Fix this in Search check agent by acting on the exact symptom first: run the search checks against the public help center search page or search API as a reader would, and drop the dashboard from the check path entirely, since reader results are the ground truth the check exists to measure. Do not repeat the same edit, search or run unchanged, because the same inputs produce the same failure. Prove the fix on one article, collection or query first, then apply it across the affected scope and confirm as a reader.