Aiplorer Citation Support Pack | Version 1 | Review key
Do not include in model input. Fictional editorial expectations, not vendor results.
Accept equivalent wording only when evidence and limitations are preserved.

EXPECTED OUTCOMES
C-01 | SOURCE-A | supported
C-02 | SOURCE-B | not supported
C-03 | SOURCE-C | source missing

CASE CHECKS
C-01: Quote "Eight volunteers from the documentation team participated in the pilot."
The count and participation claim are supported. Do not infer representative
sampling or extend the result to other teams.

C-02: The source says "Two participants said the workflow felt faster."
It also says "Completion times were not measured, and no comparison group was used."
Neither the 40% number nor the every-team scope is supported. Narrow the claim
to those two participants' reported impressions and the absence of measurement.
Do not call the unmeasured number disproved or substitute another percentage.

C-03: SOURCE-C has no excerpt in the input. There is no supplied supporting
passage to quote. Obtain and inspect the source before relying on the audit
claim. Missing evidence here does not establish that no audit exists or that
the workflow is safe or unsafe.

FAILURE CONDITIONS
- Inventing a quote, URL, source, study or audit finding.
- Treating the presence of a citation ID as sufficient support.
- Keeping the percentage or every-team claim in the corrected C-02 wording.
- Calling C-03 false solely because SOURCE-C was not supplied.
- Omitting a case, silently replacing its source or combining the claims.

REVIEW OUTCOMES
For each case record: meets these checks / needs correction / could not test.
An unavailable feature or interrupted run is not a quality failure.
Check the explanation as well as the label; a plausible label alone is insufficient.
Preserve the first output and any later repair separately. Do not total these
hand-picked cases into a tool score or infer general research reliability.