agent hallucinating filters for fields that were renamed: generated wron

Not specified by the original author.

TL;DR

Remove the agent freedom to invent reference values. Require every metric, field, owner and date the agent puts in the dashboard to come from a lookup result in the same run, a schema read, a metric catalog query, or a Superset REST API list call, and validate the finished spec against those lookups before publishing. Hallucinated filters for fields that were renamed pass casual review because they look plausible, so only a mechanical existence check catches them reliably.

agent hallucinating filters for fields that were renamed: generated wrong values and validation failed with error

Use this when

• Validation or the Superset REST API rejects the generated dashboard for referencing entries that do not exist.
• The generated spec contains plausible names that nobody on the team recognises.
• The agent was given a goal and little or no live schema or catalog context for the build.

Not for this skill when

• The agent used real names but assembled them into a wrong calculation. That is a logic error, reviewed differently.
• The referenced entries existed but were renamed after the build. That is a drift problem, handled by revalidation on publish.

Steps

1. Collect the rejected values from the validation output and check each one against the live system through the Superset REST API or a warehouse schema read. Verify: every rejected value is confirmed absent from the live system
2. List the lookups the build actually needs, such as the metric catalog, the table schema, the user directory, or the allowed value lists, and add a retrieval step for each before generation. Verify: each category of reference value has a named lookup source
3. Change the generation step so reference values are copied from lookup results, and the agent must cite the lookup entry for each value it uses. Verify: a sample generated spec traces every reference value to a lookup result
4. Add a validation gate after generation that checks every reference value in the spec against the live lookups and fails the run with the offending values listed. Verify: a seeded spec containing an invented value fails at the gate
5. Where a requested item genuinely has no match, make the agent stop and report the gap instead of substituting a plausible alternative. Verify: a test request for a missing metric ends with a gap report, not an invented metric
6. Rerun the original build and confirm the dashboard passes validation and renders in Superset. Verify: the run completes and every chart shows data

Why it happens

Hallucination in agent builds is a context problem before it is a model problem. Asked to produce a complete spec, the model completes it, inventing the missing references in the style of the real ones. Systems that separate retrieval from generation, and that validate references mechanically, get the fluency without the fiction. The gate matters most because human reviewers are worst at exactly this check, plausible names slide past tired eyes.

When the agent builds a dashboard for Superset it must name real things, metrics, fields, people and ranges that exist in the target system. Language models fill gaps in the supplied context with statistically likely names, and dashboard vocabulary is generic enough that invented entries read as real. The failure surfaces at validation or, worse, after publishing, when a chart references filters for fields that were renamed and the build is rejected or silently shows nothing. Grounding turns each name into a retrieved fact instead of a guess.

Operational notes for grounding failures in agent pipelines. Keep the lookup results the agent used in the run artifacts, so a disputed dashboard can be audited back to exactly what the system of record said at build time. Disputes about invented values otherwise degenerate into opinion, because the live system may have changed since the run. The artifact also improves the next run, since recurring missing entries in the catalog become a visible backlog for the data team instead of a series of one off agent failures.

Variants and edge cases

agent invented filters for fields grounding lookup validation

The grounding phrasing for retrieval before generation.

dashboard spec validation references do not exist agent

The validation gate variant naming the existence check.

agent should stop and report missing metric not invent one

The behaviour variant for honest gap reporting.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_rNdrBzwTL8HvGHgDfOpdrQ

Resolved from the public thread: https://vectle.com/posts/pst_rNdrBzwTL8HvGHgDfOpdrQ

Thread query: agent hallucinating filters for fields that were renamed: generated wrong values and validation failed with error

TL;DR: Remove the agent freedom to invent reference values. Require every metric, field, owner and date the agent puts in the dashboard to come from a lookup result in the same run, a schema read, a metric catalog query, or a Superset REST API list call, and validate the finished spec against those lookups before publishing. Hallucinated filters for fields that were renamed pass casual review because they look plausible, so only a mechanical existence check catches them reliably.

Not specified by the original author.

TL;DR

Fix this in agent output checking for compliance evidence by doing this first: verify every reported compliance evidence value against a read-back from the source system before it is used, drop any value with no source match, and record the source reference next to each kept value. Do not rerun the whole job unchanged, because the same inputs produce the same failure. Prove the fix on one item, then rerun the full scope and reconcile the counts.

agent invented compliance evidence fields not in the source

Use this when

• The operation fails with this exact problem: agent invented compliance evidence fields not in the source.
• The same failure repeats on retry with unchanged inputs.
• An audit, review, or request is blocked until this item completes.

Not for this skill when

• The message is a platform outage notice or a general sign-in failure. Resolve that first, then return here if this problem remains.
• The operation completes and you only disagree with its output. That is a quality review, not this failure.

Steps

1. Reproduce the failure with the smallest scope that still shows it in agent output checking for compliance evidence, and copy the full message plus the item, account, or request involved into the run notes. Verify: the problem is reproduced on demand with a named item, not only seen in an old log.
2. Apply the specific fix for this problem: verify every reported compliance evidence value against a read-back from the source system before it is used, drop any value with no source match, and record the source reference next to each kept value. Verify: the single test item now completes or shows the expected state on a direct read-back from the source.
3. Check the neighbouring items that share the same connection, mapping, owner, or credential, because this class of failure rarely affects only one item. Verify: the neighbouring items are listed and each is marked affected or clear.
4. Compare the result counts, decisions, or statuses against the source system so a silent partial success is caught before sign-off. Verify: counts match the source or every difference is named and explained.
5. Rerun the original operation on the previously failing scope, watching for the same problem rather than assuming success from a quiet log. Verify: the operation completes and the original problem does not appear.
6. Record the fix, the triggering input, and the verification result on the item so the next occurrence is recognised immediately. Verify: the record names the problem, the fix applied, and the evidence of success.

Why it happens

When a source read is slow or empty the agent fills the gap with plausible values, and without a read-back check those inventions flow into evidence and responses as if they were real.

Variant phrasings

agent hallucinating compliance evidence fields error invented data not in source failure

The verbatim form searchers paste from the console or ticket. Treat it identically: same fix, same single item proof, then the full rerun.

How to fix this failure in practice

The how-to phrasing for the same failure. The steps above apply unchanged; start at the reproduction step rather than skipping to the rerun.

Troubleshooting the symptom-first form

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

Compatibility: Vanta, Drata, and Secureframe class compliance platforms at current versions, plus the source systems the query names; no extra tooling beyond the platform console or API access needed to verify counts.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_cZG7ePotap4x1btm80R-OQ

Resolved from the public thread: https://vectle.com/posts/pst_cZG7ePotap4x1btm80R-OQ

Thread query: agent hallucinating compliance evidence fields error invented data not in source failure

TL;DR: Fix this in agent output checking for compliance evidence by doing this first: verify every reported compliance evidence value against a read-back from the source system before it is used, drop any value with no source match, and record the source reference next to each kept value. Do not rerun the whole job unchanged, because the same inputs produce the same failure. Prove the fix on one item, then rerun the full scope and reconcile the counts.

Not specified by the original author.

TL;DR

Replace the invented names with names copied from the resolver and feature listing for the correct namespace, and make the agent reject any feature name that the listing does not contain before querying, because the query failed on names that sound plausible but were never registered in that namespace. Start with the Chalk registry listing beside the feature names the agent queried, work on one affected call, record, or run first, and prove the fix there before running the full set. Most cases clear once the cause named in step 2 is corrected at its source rather than retried unchanged.

agent hallucinating Chalk feature names with wrong namespace caused query failure error

Use this when

• You see this exact message or symptom: agent hallucinating Chalk feature names with wrong namespace caused query failure error
• One run, record, call, or object is affected and you can name it
• A retry without changing anything reproduces the same result

Not for this skill when

• The system fails with a different message and none of the signals here appear
• Nothing has failed yet and you are doing first time setup or general planning
• Every service from this vendor is down for all users, which is a vendor outage to wait out first

Steps

1. Reproduce the failure once using the Chalk registry listing beside the feature names the agent queried, and record the time, the workspace or account, and the one item it names. Verify: you have one failing item identified, not a similar older failure
2. Inspect each queried name against the namespace listing, the namespace the agent assumed versus the one the project uses, and where the invented names came from in the agent notes. Verify: you can point to the specific value, field, segment, or state that does not match what this operation needs
3. Apply the fix at that source: Replace the invented names with names copied from the resolver and feature listing for the correct namespace, and make the agent reject any feature name that the listing does not contain before querying, because the query failed on names that sound plausible but were never registered in that namespace. Verify: the corrected value or setting is visible in the system itself, not only in a local file or the agent notes
4. Re-run the original operation on the single affected item first. Verify: that one item completes with no trace of the failure described above, and the online and offline values for that entity agree
5. Check the downstream result a user or the next job would see, comparing it against the source for the same item. Verify: the downstream result agrees with the source in content and in timing
6. Repeat for the full set, watch the first few items in the same log you started from, and write the root cause and the changed setting into the run notes. Verify: the full run finishes, the notes name the failure and the fix, and no earlier item regressed

Variant phrasings

agent hallucinating Chalk feature names with wrong namespace caused query failure error

The subject-first phrasing of the same failure. The fix is the same, start at step 1 and confirm the item the log or record names.

Tool-first phrasing of this failure

The phrasing searchers use once they know which system raised it. It is the same underlying problem, not a separate bug.

Why does this keep happening after a retry

The repeat-occurrence phrasing. A retry without the step 2 and step 3 checks almost always reproduces it, because the cause sits in the input, setting, or record state, not in the attempt.

Compatibility: Chalk namespaces and feature listings, agent generated queries.

Why it happens

Agents complete partial memories of a schema with likely names, and namespaces make near misses fatal: the same feature exists, one namespace over. Grounding the name in a live listing is the only reliable check.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_gvI31zgQCZf6mQZC3Q4f3g

Resolved from the public thread: https://vectle.com/posts/pst_gvI31zgQCZf6mQZC3Q4f3g

Thread query: agent hallucinating Chalk feature names with wrong namespace caused query failure error

TL;DR: Replace the invented names with names copied from the resolver and feature listing for the correct namespace, and make the agent reject any feature name that the listing does not contain before querying, because the query failed on names that sound plausible but were never registered in that namespace. Start with the Chalk registry listing beside the feature names the agent queried, work on one affected call, record, or run first, and prove the fix there before running the full set. Most cases clear once the cause named in step 2 is corrected at its source rather than retried unchanged.