Structured Intake Normalization
Converts forms, briefs, tickets, and unstructured requests into a validated schema; flags missing mandatory data.
REFERENCE_ONLYConverts forms, briefs, tickets, and unstructured requests into a validated schema; flags missing mandatory data.
REFERENCE_ONLYBreaks approved work into bounded tasks, dependencies, owners, gates, and deadlines.
REFERENCE_ONLYValidates lifecycle state, output completeness, provenance, next owner, and blockers before handoff.
REFERENCE_ONLYSelects the correct named approver from an approved authority matrix; never fabricates authority.
REFERENCE_ONLYExtracts operational controls, prohibitions, exceptions, and review triggers from approved policy texts.
REFERENCE_ONLYExtracts material factual claims and attaches traceable sources, dates, context, and uncertainty.
REFERENCE_ONLYAssesses authority, primary/secondary status, recency, relevance, conflict, and provenance of evidence.
REFERENCE_ONLYRetrieves approved information and synthesizes it without inventing unsupported facts; exposes source gaps.
REFERENCE_ONLYApplies controlled vocabularies, entity IDs, relationships, classification labels, and data-quality flags.
REFERENCE_ONLYRecords source assets, transformations, prompts/workflows, contributors, tools, rights, and derivatives.
REFERENCE_ONLYIdentifies probable personal, confidential, or restricted data and produces minimized/redacted derivative views.
REFERENCE_ONLYAssigns risk tier based on defined policy signals and routes high/restricted situations to accountable humans.
REFERENCE_ONLYConverts approved outcomes into testable, observable requirements, constraints, and acceptance conditions.
REFERENCE_ONLYMaps inputs, decisions, controls, roles, steps, exceptions, and measurable outcomes.
REFERENCE_ONLYCompares alternatives against explicit criteria, uncertainty, cost, risk, reversibility, and constraints.
REFERENCE_ONLYOrganizes interviews, open-text feedback, and observation notes into auditable themes, quotes, contradictions, and evidence strength.
REFERENCE_ONLYProduces reproducible summaries, visualizations, segmentation, anomalies, and interpretation with stated limits.
REFERENCE_ONLYCreates transparent base/upside/downside forecasts with explicit assumptions, sensitivity, and backtesting where applicable.
REFERENCE_ONLYDefines hypothesis, sample, assignments, metrics, guardrails, stopping rules, and cautious readouts.
REFERENCE_ONLYDrafts audience-appropriate content, documents, messages, and summaries from approved sources and tone constraints.
REFERENCE_ONLYCompares a draft to approved voice, personality, visual/sonic standards, and anti-examples; identifies drift.
REFERENCE_ONLYFinds suspicious semantic, phrasing, visual, or sonic similarity and imitation cues; provides evidence, not legal conclusions.
REFERENCE_ONLYChecks readability, captions, contrast/design guidance, keyboard/flow considerations, alt-text, and inclusive language.
REFERENCE_ONLYEvaluates visual assets against approved references, layout systems, character rules, artifacts, safe areas, and format requirements.
REFERENCE_ONLYTests speech/music continuity, pronunciation, artifacts, loudness, metadata, and technical readiness.
REFERENCE_ONLYCreates maintainable code and tests within a bounded approved specification, including safe dependency and secret handling.
REFERENCE_ONLYIdentifies correctness, maintainability, test, security, dependency, and policy risks in change sets.
REFERENCE_ONLYCreates positive, negative, boundary, regression, integration, and recovery tests; records reproducible evidence.
REFERENCE_ONLYCorrelates approved telemetry, assesses scope/severity by policy, retrieves runbooks, and routes escalation.
REFERENCE_ONLYIdentifies attack surfaces, threat scenarios, compensating controls, evidence, and remediation priority.
REFERENCE_ONLYMeasures completeness, validity, freshness, duplicate rates, distributions, and schema drift without silently editing data.
REFERENCE_ONLYTests AI systems for task quality, grounding, leakage, unsafe behavior, bias risk, regressions, and refusal behavior.
REFERENCE_ONLYStructures clauses, dates, parties, obligations, limitations, and comparison flags without interpreting law.
REFERENCE_ONLYMatches approved records, detects discrepancies, prioritizes exceptions, and prepares support evidence.
REFERENCE_ONLYMaintains risk statements, owners, indicators, mitigations, evidence, and due dates; flags overdue/residual risk.
REFERENCE_ONLYClassifies requests, retrieves approved guidance, drafts noncommittal responses, and detects urgent/sensitive escalation.
REFERENCE_ONLYMaps approved campaign objectives to audiences, content pillars, assets, sequencing, test hypotheses, and measurement.
REFERENCE_ONLYDetects abuse, threats, self-harm indicators, harassment, vulnerable audiences, and reputation-sensitive conversations.
REFERENCE_ONLYMaps source/provider terms, permissions, expiry, territories, restrictions, and derivative eligibility.
REFERENCE_ONLYVerifies final assets, metadata, format, checksums, approvals, evidence, and delivery status against a declared checklist.
REFERENCE_ONLYProduces capacity-aware schedule options and collision/constraint warnings from authorized data.
REFERENCE_ONLYOrganizes nonconformance evidence, root-cause hypotheses, containment drafts, corrective actions, and verification criteria.
REFERENCE_ONLYAdapts content for language/region while preserving meaning, voice, rights, and local requirements; flags unverifiable claims.
REFERENCE_ONLYConverts conversations into clearly attributed decisions, action items, owners, deadlines, open questions, and risks.
REFERENCE_ONLYApplies rubric-based quality, policy, technical, and completeness checks to produce reproducible pass/fail evidence.
REFERENCE_ONLY