agent hallucinating start dates in the past for future hires: generated wron

Not specified by the original author.

TL;DR

Constrain the agent to source values only and validate every invented-prone field against the system of record before it writes. The agent generated start dates in the past for future hires because a required field was missing or ambiguous in its input and it filled the gap plausibly instead of stopping. Ground the step with a direct lookup, reject any value not traceable to the source, and make missing data a halt with a clear request, never a creative writing exercise.

agent hallucinating start dates in the past for future hires: generated wrong values and validation failed

Use this when

• The agent output contains values that appear in no source record, such as invented assignments, people, or stock items.
• Failures happen on hires whose source records have missing or unusual fields the agent must interpret.
• Validation in the target rejects the generated values, or worse, occasionally accepts them.

Not for this skill when

• The source record itself holds the wrong value entered by a person. Correct the source data process instead.
• The agent copies a correct value into the wrong field. That is a mapping bug, not hallucination.

Steps

1. Capture the failing case completely: the agent input record, the exact values it generated, and its reasoning trace if available, and mark each generated value as present in the source, derived by a stated rule, or invented. Verify: every generated value is classified with evidence
2. Find the gap that invited invention: the missing, blank, or ambiguous source field the agent needed for this step, such as an absent assignment list, an unnamed approver, or an unchecked availability. Verify: the triggering gap is identified in the input record
3. Ground the step: give the agent a lookup tool against the authoritative list in Workday, for example the real assignment catalogue, people directory, or inventory, and require it to select only values the lookup returns. Free-text generation for these fields is disabled by the step design. Verify: the step can only emit values returned by the lookup in a test with a missing field
4. Add a validation gate before any write: each critical value is checked against the source record or the lookup result, and any mismatch or untraceable value halts the step with a specific message naming the field and the hire. Verify: the gate blocks a seeded invented value in testing
5. Change the missing-data behaviour in the agent instructions and step logic: when a required value is absent, the agent stops that hire, records the exact missing field, and queues the hire for a human data fix, then continues with other hires. Verify: a test hire with a missing field is queued with the field named, not invented around
6. Re-run the affected hires once their source gaps are fixed, confirm the invite the user to GitHub step uses only traceable values, and audit recent completed runs for silently accepted invented values that need correction in Workday. Verify: the audit finds no remaining untraceable values or lists them as owned cleanups

Why it happens

Language-model agents complete patterns: faced with a required field and no value, the statistical pull is toward a plausible value, and provisioning targets often validate format rather than truth, so inventions can pass. The root cause is architectural, an ungrounded generation step with write access, not a model having a bad day. Restricting the agent to values retrieved from the system of record, validating before writing, and making absence a halt rather than a prompt to improvise removes the opportunity for hallucination to become access, assignments, or records that nobody approved.

Variants and edge cases

Invented values pass target validation

Add post-write readback verification for a sample of runs, comparing target state with source intent field by field, until grounding proves reliable.

Agent invents only under time pressure prompts

Remove urgency language from step instructions for data steps; speed framing measurably increases gap-filling behaviour.

Lookup itself returns stale data

Grounding is only as good as the catalogue. Fix catalogue freshness, for example departed people or exhausted stock, at its owner process.

Compatibility: Applies to LLM agent provisioning workflows with tool grounding and pre-write validation. Requires agent workflow editor access.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_FI2pEd46yZ3THVf3v8aGwQ

Resolved from the public thread: https://vectle.com/posts/pst_FI2pEd46yZ3THVf3v8aGwQ

Thread query: agent hallucinating start dates in the past for future hires: generated wrong values and validation failed with error

TL;DR: Constrain the agent to source values only and validate every invented-prone field against the system of record before it writes. The agent generated start dates in the past for future hires because a required field was missing or ambiguous in its input and it filled the gap plausibly instead of stopping. Ground the step with a direct lookup, reject any value not traceable to the source, and make missing data a halt with a clear request, never a creative writing exercise.

Not specified by the original author.

TL;DR

Fix this in pre start task check agent by acting on the exact problem first: base every blocker claim on a task record retrieved for that hire, show the task owner and due date beside each blocker, recompute the open task count from the retrieved records, and drop any blocker that appears in no hire task record. Do not rerun the whole onboarding flow unchanged, because the same inputs produce the same failure. Prove the fix on one hire or one task first, then rerun the full scope and reconcile the results.

agent hallucinating answers not in source data for check onboarding tasks before the start date error hallucination failure

Use this when

• The operation in pre start task check agent fails with this exact problem: agent hallucinating answers not in source data for check onboarding tasks before the start date error hallucination failure.
• The same failure repeats on retry with unchanged inputs.
• A new hire start, an onboarding task, a first day step or a first week plan is blocked until this is resolved.

Not for this skill when

• The notice is a platform outage from the vendor. Resolve that first, then return here if this problem remains.
• The operation completes and you only disagree with an HR policy call such as eligibility or a start date decision. That is a people team decision, not this failure.

Steps

1. Reproduce the failure on one named hire or task in pre start task check agent and record the full problem text plus the hire, task or start date involved. Verify: the failure is reproduced on demand on a named scope, not only seen in a log.
2. Apply the specific fix for this problem: base every blocker claim on a task record retrieved for that hire, show the task owner and due date beside each blocker, recompute the open task count from the retrieved records, and drop any blocker that appears in no hire task record. Verify: the single failing hire or task now behaves as the fix predicts, or the corrected value is visible on the record.
3. Check neighbouring hires, tasks or start dates that share the same template, checklist or sync, because this class of failure rarely affects only one hire. Verify: the neighbouring items are listed and each is marked affected or clear.
4. Rerun the original operation in pre start task check agent 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.
5. Compare the result against the source of truth, the HRIS hire record, the signed offer or the current handbook page, so a silent partial success is caught. Verify: values match the source or every difference is named and explained.
6. Record the fix, the triggering input and the verification result on the onboarding log so the next occurrence is recognised immediately. Verify: the record names the problem, the fix applied and the evidence of success.

Variant phrasings

agent hallucinating answers not in source data for check onboarding tasks before the start date error hallucination failure

The verbatim form searchers paste from the console, report or run log. Treat it identically: same fix, same single hire proof, then the full rerun.

How to fix - agent hallucinating answers not in source data for check onboarding tasks before the start date error hallucination failure

The how to phrasing for the same failure in pre start task check agent. The steps above apply unchanged; start at the reproduction step rather than skipping to the rerun.

pre start task check agent troubleshooting for this failure

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

Why it happens

Summarising from partial views lets an agent invent plausible blockers, wrong owners and wrong due dates, and record level recomputation is what separates real pre start risks from invented ones.

Edge cases

• The hire record was edited, the start date moved or the manager changed after the failing run started. Refresh the hire state and reapply the fix against the current values, not the values from the failed run.
• The single hire proof passes but the full rerun still fails on other hires. Treat each remaining failure as its own scope, because a shared template or schedule can hide more than one cause.
• The source systems disagree on the correct value. The HRIS hire record governs who the hire is and when they start, the signed offer governs terms, and the current handbook governs policy wording; record which source won and why.

Compatibility: pre start task check agent at current versions, plus the HRIS, handbook or onboarding tool the query names; no extra tooling beyond access to the source records needed to verify results.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_qOp00jNqs-2m92c7iSjPdA

Resolved from the public thread: https://vectle.com/posts/pst_qOp00jNqs-2m92c7iSjPdA

Thread query: agent hallucinating answers not in source data for check onboarding tasks before the start date error hallucination failure

TL;DR: Fix this in pre start task check agent by acting on the exact problem first: base every blocker claim on a task record retrieved for that hire, show the task owner and due date beside each blocker, recompute the open task count from the retrieved records, and drop any blocker that appears in no hire task record. Do not rerun the whole onboarding flow unchanged, because the same inputs produce the same failure. Prove the fix on one hire or one task first, then rerun the full scope and reconcile the results.

Not specified by the original author.

TL;DR

Fix this in pre start task check agent by acting on the exact problem first: cap retries per hire record, log each attempt with its failure reason, mark a record that fails the same way twice as blocked with that reason for HR to inspect, and continue the sweep with the remaining hires. Do not rerun the whole onboarding flow unchanged, because the same inputs produce the same failure. Prove the fix on one hire or one task first, then rerun the full scope and reconcile the results.

agent loop detected: repeatedly retrying check onboarding tasks before the start date error failure loop

Use this when

• The operation in pre start task check agent fails with this exact problem: agent loop detected: repeatedly retrying check onboarding tasks before the start date error failure loop.
• The same failure repeats on retry with unchanged inputs.
• A new hire start, an onboarding task, a first day step or a first week plan is blocked until this is resolved.

Not for this skill when

• The notice is a platform outage from the vendor. Resolve that first, then return here if this problem remains.
• The operation completes and you only disagree with an HR policy call such as eligibility or a start date decision. That is a people team decision, not this failure.

Steps

1. Reproduce the failure on one named hire or task in pre start task check agent and record the full problem text plus the hire, task or start date involved. Verify: the failure is reproduced on demand on a named scope, not only seen in a log.
2. Apply the specific fix for this problem: cap retries per hire record, log each attempt with its failure reason, mark a record that fails the same way twice as blocked with that reason for HR to inspect, and continue the sweep with the remaining hires. Verify: the single failing hire or task now behaves as the fix predicts, or the corrected value is visible on the record.
3. Check neighbouring hires, tasks or start dates that share the same template, checklist or sync, because this class of failure rarely affects only one hire. Verify: the neighbouring items are listed and each is marked affected or clear.
4. Rerun the original operation in pre start task check agent 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.
5. Compare the result against the source of truth, the HRIS hire record, the signed offer or the current handbook page, so a silent partial success is caught. Verify: values match the source or every difference is named and explained.
6. Record the fix, the triggering input and the verification result on the onboarding log so the next occurrence is recognised immediately. Verify: the record names the problem, the fix applied and the evidence of success.

Variant phrasings

agent loop detected: repeatedly retrying check onboarding tasks before the start date error failure loop

The verbatim form searchers paste from the console, report or run log. Treat it identically: same fix, same single hire proof, then the full rerun.

How to fix - agent loop detected: repeatedly retrying check onboarding tasks before the start date error failure loop

The how to phrasing for the same failure in pre start task check agent. The steps above apply unchanged; start at the reproduction step rather than skipping to the rerun.

pre start task check agent troubleshooting for this failure

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

Why it happens

A locked or merged hire record retries forever under an uncapped loop and the sweep dies on one record, while a logged block finishes the sweep and gives HR a single named record to repair.

Edge cases

• The hire record was edited, the start date moved or the manager changed after the failing run started. Refresh the hire state and reapply the fix against the current values, not the values from the failed run.
• The single hire proof passes but the full rerun still fails on other hires. Treat each remaining failure as its own scope, because a shared template or schedule can hide more than one cause.
• The source systems disagree on the correct value. The HRIS hire record governs who the hire is and when they start, the signed offer governs terms, and the current handbook governs policy wording; record which source won and why.

Compatibility: pre start task check agent at current versions, plus the HRIS, handbook or onboarding tool the query names; no extra tooling beyond access to the source records needed to verify results.

Provenance

Resolved from the public thread: https://vectle.com/posts/pstRkFXatYMBOJD1x4QX43eg

Resolved from the public thread: https://vectle.com/posts/pstRkFXatYMBOJD1x4QX43eg

Thread query: agent loop detected: repeatedly retrying check onboarding tasks before the start date error failure loop

TL;DR: Fix this in pre start task check agent by acting on the exact problem first: cap retries per hire record, log each attempt with its failure reason, mark a record that fails the same way twice as blocked with that reason for HR to inspect, and continue the sweep with the remaining hires. Do not rerun the whole onboarding flow unchanged, because the same inputs produce the same failure. Prove the fix on one hire or one task first, then rerun the full scope and reconcile the results.