Define the exact schema
Name each field, explain its meaning, choose its type, and mark whether it is required.
data.extractUSE CASE 02 / STRUCTURED DATA
EXTRACT.
THEN PROVE.
The calling agent defines the fields. ActionDock extracts only what the supplied text supports, attaches evidence, reports missing values, and applies explicit rules.
01SCHEMADefine fields and types
02EXTRACTReturn values with evidence
03VALIDATEApply deterministic rules
04ROUTELet the agent choose next work
JSON WITH PROVENANCE
Missing means
missing.
Absent source data is not guessed. Every requested field returns a value and evidence pair, with required gaps listed separately.
Preserve field evidence
Each returned value includes a short supporting excerpt or null when the source does not support it.
get_jobApply explicit rules
Check required values, types, equality, allowed sets, containment, prefixes, and numeric or length bounds.
data.validateRoute only valid records
The agent decides whether to request missing information, create a report, or propose a supervised system update.
report.create · integration.executeBUSINESS PATTERNS
Structure the work that arrives as text.
VENDOR INTAKE
Extract ownership, security, renewal, and service fields before a procurement review.
REQUEST TRIAGE
Turn an operational request into a complete record and flag missing required context.
DATA QUALITY
Validate records read from an approved system and produce a precise correction report.
CALLER-DEFINED SCHEMA