Capability library · 141 skills
Every skill has a stage, a purpose, and evidence.
Search traditional, augmented, anchor, and cross-cutting capabilities. Every detail page links to its stage, agents, source documents, and local Markdown file.
Stakeholder interviewing
Stakeholder interviewing is an established design capability used to turn stakeholder intent, constraints, and conflict into an accountable charter. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Business analysis
Business analysis is an established design capability used to turn stakeholder intent, constraints, and conflict into an accountable charter. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Requirements documentation
Requirements documentation is an established design capability used to turn stakeholder intent, constraints, and conflict into an accountable charter. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Facilitation
Facilitation is an established design capability used to turn stakeholder intent, constraints, and conflict into an accountable charter. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Scope negotiation
Scope negotiation is an established design capability used to turn stakeholder intent, constraints, and conflict into an accountable charter. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
AI-assisted interview synthesis
AI-assisted interview synthesis helps a designer turn stakeholder intent, constraints, and conflict into an accountable charter. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Async interview design
Async interview design is the discipline of structuring questions for an AI moderator to ask adaptively. It helps a designer turn stakeholder intent, constraints, and conflict into an accountable charter while keeping the final judgment traceable and human-owned.
Voice/text transcription review
Voice/text transcription review helps a designer turn stakeholder intent, constraints, and conflict into an accountable charter. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Requirement clustering
Requirement clustering helps a designer turn stakeholder intent, constraints, and conflict into an accountable charter. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Conflict-flagging interpretation
Conflict-flagging interpretation helps a designer turn stakeholder intent, constraints, and conflict into an accountable charter. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
PRD/charter drafting from AI synthesis
PRD/charter drafting from AI synthesis helps a designer turn stakeholder intent, constraints, and conflict into an accountable charter. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Business-constraint extraction
Business-constraint extraction helps a designer turn stakeholder intent, constraints, and conflict into an accountable charter. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Prompt engineering for stakeholder intake
Prompt engineering for stakeholder intake helps a designer turn stakeholder intent, constraints, and conflict into an accountable charter. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
AI-brief validation
AI-brief validation is the discipline of checking a drafted charter against what was actually said, not just what reads well. It helps a designer turn stakeholder intent, constraints, and conflict into an accountable charter while keeping the final judgment traceable and human-owned.
Ethnographic research
Ethnographic research is an established design capability used to build a verified fact base without confusing generated synthesis with lived evidence. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Contextual inquiry
Contextual inquiry is an established design capability used to build a verified fact base without confusing generated synthesis with lived evidence. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Survey design
Survey design is an established design capability used to build a verified fact base without confusing generated synthesis with lived evidence. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Competitive analysis
Competitive analysis is an established design capability used to build a verified fact base without confusing generated synthesis with lived evidence. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Analytics review
Analytics review is an established design capability used to build a verified fact base without confusing generated synthesis with lived evidence. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Interview moderation
Interview moderation is an established design capability used to build a verified fact base without confusing generated synthesis with lived evidence. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
AI-assisted feedback clustering
AI-assisted feedback clustering helps a designer build a verified fact base without confusing generated synthesis with lived evidence. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Outlier-flagging review
Outlier-flagging review is the discipline of confirming the agent flagged, not discarded, anomalies. It helps a designer build a verified fact base without confusing generated synthesis with lived evidence while keeping the final judgment traceable and human-owned.
Persona synthesis from AI drafts
Persona synthesis from AI drafts helps a designer build a verified fact base without confusing generated synthesis with lived evidence. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
JTBD statement construction
JTBD statement construction helps a designer build a verified fact base without confusing generated synthesis with lived evidence. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Persona validation against live research
Persona validation against live research helps a designer build a verified fact base without confusing generated synthesis with lived evidence. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Synthetic/AI-moderated interview design
Synthetic/AI-moderated interview design helps a designer build a verified fact base without confusing generated synthesis with lived evidence. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Synthetic-vs-real calibration
Synthetic-vs-real calibration is the discipline of knowing when synthetic data is and isn't a valid proxy. It helps a designer build a verified fact base without confusing generated synthesis with lived evidence while keeping the final judgment traceable and human-owned.
Competitor crawl configuration
Competitor crawl configuration helps a designer build a verified fact base without confusing generated synthesis with lived evidence. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Feature-matrix interpretation
Feature-matrix interpretation helps a designer build a verified fact base without confusing generated synthesis with lived evidence. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Research-repository tagging/synthesis
Research-repository tagging/synthesis is the discipline of Dovetail/Condens-class tools. It helps a designer build a verified fact base without confusing generated synthesis with lived evidence while keeping the final judgment traceable and human-owned.
Ethical data curation
Ethical data curation helps a designer build a verified fact base without confusing generated synthesis with lived evidence. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Cross-source evidence triangulation
Cross-source evidence triangulation helps a designer build a verified fact base without confusing generated synthesis with lived evidence. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Strategic-differentiation judgment
Strategic-differentiation judgment helps a designer build a verified fact base without confusing generated synthesis with lived evidence. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Cultural/behavioral pattern interpretation
Cultural/behavioral pattern interpretation helps a designer build a verified fact base without confusing generated synthesis with lived evidence. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Synthesis/affinity mapping
Synthesis/affinity mapping is an established design capability used to convert evidence into a specific, falsifiable problem worth solving. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
"How Might We" writing
"How Might We" writing is an established design capability used to convert evidence into a specific, falsifiable problem worth solving. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Problem-statement authorship
Problem-statement authorship is an established design capability used to convert evidence into a specific, falsifiable problem worth solving. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Workshop facilitation for framing
Workshop facilitation for framing is an established design capability used to convert evidence into a specific, falsifiable problem worth solving. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Candidate-framing generation review
Candidate-framing generation review is the discipline of evaluating an AI-drafted set of problem statements. It helps a designer convert evidence into a specific, falsifiable problem worth solving while keeping the final judgment traceable and human-owned.
Evidence-traceability checking
Evidence-traceability checking is the discipline of does this framing actually cite the data. It helps a designer convert evidence into a specific, falsifiable problem worth solving while keeping the final judgment traceable and human-owned.
Falsifiability testing
Falsifiability testing is the discipline of stress-testing a statement against contradicting evidence. It helps a designer convert evidence into a specific, falsifiable problem worth solving while keeping the final judgment traceable and human-owned.
Framing-selection judgment
Framing-selection judgment is the discipline of choosing fit over polish. It helps a designer convert evidence into a specific, falsifiable problem worth solving while keeping the final judgment traceable and human-owned.
Prompt engineering for structured evidence synthesis
Prompt engineering for structured evidence synthesis helps a designer convert evidence into a specific, falsifiable problem worth solving. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Problem-statement ownership
Problem-statement ownership is the discipline of specific discipline of editing, not adopting verbatim. It helps a designer convert evidence into a specific, falsifiable problem worth solving while keeping the final judgment traceable and human-owned.
Sketching/ideation facilitation
Sketching/ideation facilitation is an established design capability used to expand the solution space while preserving taste, ethics, originality, and strategic fit. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Applying innovation frameworks
Applying innovation frameworks is an established design capability used to expand the solution space while preserving taste, ethics, originality, and strategic fit. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Creative critique
Creative critique is an established design capability used to expand the solution space while preserving taste, ethics, originality, and strategic fit. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Brand-fit judgment
Brand-fit judgment is an established design capability used to expand the solution space while preserving taste, ethics, originality, and strategic fit. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Concept selection
Concept selection is an established design capability used to expand the solution space while preserving taste, ethics, originality, and strategic fit. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Divergent-prompt design
Divergent-prompt design is the discipline of getting genuine structural variety, not repetition, out of an ideation agent. It helps a designer expand the solution space while preserving taste, ethics, originality, and strategic fit while keeping the final judgment traceable and human-owned.
Anti-solution interpretation
Anti-solution interpretation helps a designer expand the solution space while preserving taste, ethics, originality, and strategic fit. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Volume-to-signal triage
Volume-to-signal triage is the discipline of fast filtering of large AI-generated concept sets. It helps a designer expand the solution space while preserving taste, ethics, originality, and strategic fit while keeping the final judgment traceable and human-owned.
AI-assisted low-fi sketch review
AI-assisted low-fi sketch review helps a designer expand the solution space while preserving taste, ethics, originality, and strategic fit. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Concept-critique prompting
Concept-critique prompting is the discipline of directing a Design Critique Agent. It helps a designer expand the solution space while preserving taste, ethics, originality, and strategic fit while keeping the final judgment traceable and human-owned.
Ethical-veto judgment
Ethical-veto judgment is the discipline of sensitive-context concepts. It helps a designer expand the solution space while preserving taste, ethics, originality, and strategic fit while keeping the final judgment traceable and human-owned.
Innovation-leap recognition
Innovation-leap recognition is the discipline of knowing where AI's interpolation stops and a real leap is needed. It helps a designer expand the solution space while preserving taste, ethics, originality, and strategic fit while keeping the final judgment traceable and human-owned.
Rationale documentation for selection decisions
Rationale documentation for selection decisions helps a designer expand the solution space while preserving taste, ethics, originality, and strategic fit. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Feasibility-flag interpretation
Feasibility-flag interpretation is the discipline of AI can't guarantee feasibility, knowing when to pull in engineering early. It helps a designer expand the solution space while preserving taste, ethics, originality, and strategic fit while keeping the final judgment traceable and human-owned.
Bias-toward-convention detection
Bias-toward-convention detection is the discipline of catching when AI output quietly narrows toward the safe/common pattern. It helps a designer expand the solution space while preserving taste, ethics, originality, and strategic fit while keeping the final judgment traceable and human-owned.
Prompt engineering for framework-driven ideation
Prompt engineering for framework-driven ideation helps a designer expand the solution space while preserving taste, ethics, originality, and strategic fit. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Sitemap/nav design
Sitemap/nav design is an established design capability used to shape information into a structure people can understand and navigate. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Taxonomy development
Taxonomy development is an established design capability used to shape information into a structure people can understand and navigate. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Content strategy
Content strategy is an established design capability used to shape information into a structure people can understand and navigate. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Cognitive-load assessment
Cognitive-load assessment is an established design capability used to shape information into a structure people can understand and navigate. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
AI-generated sitemap review
AI-generated sitemap review helps a designer shape information into a structure people can understand and navigate. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Cognitive-load metric interpretation
Cognitive-load metric interpretation is the discipline of clicks-to-goal as diagnostic, not verdict. It helps a designer shape information into a structure people can understand and navigate while keeping the final judgment traceable and human-owned.
Structural simplification
Structural simplification is the discipline of specific human discipline of removing AI-mapped steps. It helps a designer shape information into a structure people can understand and navigate while keeping the final judgment traceable and human-owned.
Localization/taxonomy adaptation
Localization/taxonomy adaptation helps a designer shape information into a structure people can understand and navigate. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
AI-drafted microcopy editing
AI-drafted microcopy editing helps a designer shape information into a structure people can understand and navigate. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
UX-writing prompt engineering
UX-writing prompt engineering helps a designer shape information into a structure people can understand and navigate. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Pattern-library alignment checking
Pattern-library alignment checking is the discipline of via Design System Intelligence-class tooling. It helps a designer shape information into a structure people can understand and navigate while keeping the final judgment traceable and human-owned.
Intuitiveness validation
Intuitiveness validation is the discipline of structure that scores well mechanically still needs a human "does this feel simple" pass. It helps a designer shape information into a structure people can understand and navigate while keeping the final judgment traceable and human-owned.
Cross-market taxonomy testing
Cross-market taxonomy testing helps a designer shape information into a structure people can understand and navigate. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Interaction design
Interaction design is an established design capability used to translate intent into coherent interaction, visual language, motion, and edge states. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Visual/UI design
Visual/UI design is an established design capability used to translate intent into coherent interaction, visual language, motion, and edge states. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Motion/choreography design
Motion/choreography design is an established design capability used to translate intent into coherent interaction, visual language, motion, and edge states. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Design-system application
Design-system application is an established design capability used to translate intent into coherent interaction, visual language, motion, and edge states. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Typography/color craft
Typography/color craft is an established design capability used to translate intent into coherent interaction, visual language, motion, and edge states. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Edge-case (error/empty/offline) design
Edge-case (error/empty/offline) design is an established design capability used to translate intent into coherent interaction, visual language, motion, and edge states. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Design-system-compliant generation review
Design-system-compliant generation review helps a designer translate intent into coherent interaction, visual language, motion, and edge states. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Styling-variation curation
Styling-variation curation is the discipline of fast filtering of AI-generated visual directions. It helps a designer translate intent into coherent interaction, visual language, motion, and edge states while keeping the final judgment traceable and human-owned.
Functional-prototype QA
Functional-prototype QA helps a designer translate intent into coherent interaction, visual language, motion, and edge states. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Micro-interaction accessibility review
Micro-interaction accessibility review is the discipline of catching AI-default carousels/auto-advance that create friction. It helps a designer translate intent into coherent interaction, visual language, motion, and edge states while keeping the final judgment traceable and human-owned.
Bespoke choreography authorship
Bespoke choreography authorship is the discipline of specific non-delegable skill, timing/easing that AI cannot originate. It helps a designer translate intent into coherent interaction, visual language, motion, and edge states while keeping the final judgment traceable and human-owned.
Edge-case completeness auditing
Edge-case completeness auditing is the discipline of confirming AI didn't default to happy-path-only. It helps a designer translate intent into coherent interaction, visual language, motion, and edge states while keeping the final judgment traceable and human-owned.
Cultural-sensitivity review of generated imagery
Cultural-sensitivity review of generated imagery helps a designer translate intent into coherent interaction, visual language, motion, and edge states. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Design-token/component drift detection
Design-token/component drift detection helps a designer translate intent into coherent interaction, visual language, motion, and edge states. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Prompt engineering for design-system-aware generation
Prompt engineering for design-system-aware generation is the discipline of Claude Design/Figma Make-class tools. It helps a designer translate intent into coherent interaction, visual language, motion, and edge states while keeping the final judgment traceable and human-owned.
Cross-agent handoff
Cross-agent handoff is the discipline of Interaction Agent output into Design System Agent input. It helps a designer translate intent into coherent interaction, visual language, motion, and edge states while keeping the final judgment traceable and human-owned.
Brand-soul curation
Brand-soul curation is the discipline of ensuring AI-generated variety still reads as one brand. It helps a designer translate intent into coherent interaction, visual language, motion, and edge states while keeping the final judgment traceable and human-owned.
Accessibility tuning for screen readers, motor, and cognitive impairments
Accessibility tuning for screen readers, motor, and cognitive impairments helps a designer translate intent into coherent interaction, visual language, motion, and edge states. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Usability-test moderation
Usability-test moderation is an established design capability used to test real experience quality and interpret why people succeed or struggle. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Body-language/tone reading
Body-language/tone reading is an established design capability used to test real experience quality and interpret why people succeed or struggle. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Test-plan design
Test-plan design is an established design capability used to test real experience quality and interpret why people succeed or struggle. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Root-cause analysis
Root-cause analysis is an established design capability used to test real experience quality and interpret why people succeed or struggle. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Findings prioritization
Findings prioritization is an established design capability used to test real experience quality and interpret why people succeed or struggle. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
WCAG audit interpretation
WCAG audit interpretation helps a designer test real experience quality and interpret why people succeed or struggle. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Heuristic-simulation review
Heuristic-simulation review helps a designer test real experience quality and interpret why people succeed or struggle. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
AI-moderated pre-interview design
AI-moderated pre-interview design helps a designer test real experience quality and interpret why people succeed or struggle. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Synthetic-user result calibration
Synthetic-user result calibration is the discipline of discipline of not over-trusting synthetic panels as a proxy. It helps a designer test real experience quality and interpret why people succeed or struggle while keeping the final judgment traceable and human-owned.
Click-path/task-time metric interpretation
Click-path/task-time metric interpretation helps a designer test real experience quality and interpret why people succeed or struggle. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Live-moderation-vs-AI-moderation triage
Live-moderation-vs-AI-moderation triage is the discipline of deciding which sessions need a human. It helps a designer test real experience quality and interpret why people succeed or struggle while keeping the final judgment traceable and human-owned.
Failure-interpretation
Failure-interpretation is the discipline of "why," not just the "that". It helps a designer test real experience quality and interpret why people succeed or struggle while keeping the final judgment traceable and human-owned.
Fix-prioritization against business/technical constraints
Fix-prioritization against business/technical constraints helps a designer test real experience quality and interpret why people succeed or struggle. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Accessibility-remediation writing
Accessibility-remediation writing helps a designer test real experience quality and interpret why people succeed or struggle. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Prompt engineering for usability-simulation agents
Prompt engineering for usability-simulation agents helps a designer test real experience quality and interpret why people succeed or struggle. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Cross-source finding reconciliation
Cross-source finding reconciliation is the discipline of AI-flagged vs. human-discovered. It helps a designer test real experience quality and interpret why people succeed or struggle while keeping the final judgment traceable and human-owned.
Design QA
Design QA is an established design capability used to carry design intent into code, specifications, and implementation decisions. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Developer collaboration
Developer collaboration is an established design capability used to carry design intent into code, specifications, and implementation decisions. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Spec documentation
Spec documentation is an established design capability used to carry design intent into code, specifications, and implementation decisions. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Design-engineering negotiation
Design-engineering negotiation is an established design capability used to carry design intent into code, specifications, and implementation decisions. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Generated-code visual QA
Generated-code visual QA helps a designer carry design intent into code, specifications, and implementation decisions. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Spacing/timing deviation detection
Spacing/timing deviation detection helps a designer carry design intent into code, specifications, and implementation decisions. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Design-to-code prompt engineering
Design-to-code prompt engineering is the discipline of Figma Make/Claude Design/MCP-bridge-class tools. It helps a designer carry design intent into code, specifications, and implementation decisions while keeping the final judgment traceable and human-owned.
Security/privacy review of AI-generated code
Security/privacy review of AI-generated code is the discipline of with elevated, not reduced, scrutiny. It helps a designer carry design intent into code, specifications, and implementation decisions while keeping the final judgment traceable and human-owned.
Ambiguity-flag interpretation
Ambiguity-flag interpretation is the discipline of where the agent had to guess because the design was underspecified. It helps a designer carry design intent into code, specifications, and implementation decisions while keeping the final judgment traceable and human-owned.
Asset-optimization review
Asset-optimization review helps a designer carry design intent into code, specifications, and implementation decisions. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Component unit-test review
Component unit-test review helps a designer carry design intent into code, specifications, and implementation decisions. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Design-backend contradiction negotiation
Design-backend contradiction negotiation helps a designer carry design intent into code, specifications, and implementation decisions. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
A/B test design
A/B test design is an established design capability used to read outcomes, investigate drift, and feed validated learning into the next cycle. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Statistical literacy
Statistical literacy is an established design capability used to read outcomes, investigate drift, and feed validated learning into the next cycle. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Dashboard analysis
Dashboard analysis is an established design capability used to read outcomes, investigate drift, and feed validated learning into the next cycle. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Cross-functional reporting
Cross-functional reporting is an established design capability used to read outcomes, investigate drift, and feed validated learning into the next cycle. In TADF it remains a human foundation for evaluating, correcting, and contextualising AI output.
Anomaly-flag triage
Anomaly-flag triage helps a designer read outcomes, investigate drift, and feed validated learning into the next cycle. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Root-cause investigation
Root-cause investigation is the discipline of distinguishing external events from design flaws. It helps a designer read outcomes, investigate drift, and feed validated learning into the next cycle while keeping the final judgment traceable and human-owned.
Engagement-Trap detection
Engagement-Trap detection is the discipline of named guardrail, recognizing metric-positive, trust-negative optimization. It helps a designer read outcomes, investigate drift, and feed validated learning into the next cycle while keeping the final judgment traceable and human-owned.
Variant-shipping judgment
Variant-shipping judgment helps a designer read outcomes, investigate drift, and feed validated learning into the next cycle. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
AI-generated learnings-digest editing
AI-generated learnings-digest editing helps a designer read outcomes, investigate drift, and feed validated learning into the next cycle. It combines AI-enabled speed with explicit evidence checks, correction, and accountable human judgment.
Feedback-loop authorship
Feedback-loop authorship is the discipline of writing a validated learning into the next cycle's Stage 01 charter, the skill that makes this a loop, not a report. It helps a designer read outcomes, investigate drift, and feed validated learning into the next cycle while keeping the final judgment traceable and human-owned.
Prompt engineering for anomaly-detection agents
Prompt engineering for anomaly-detection agents is the discipline of Quantum Metric/Contentsquare-class tools. It helps a designer read outcomes, investigate drift, and feed validated learning into the next cycle while keeping the final judgment traceable and human-owned.
Calibration review
Calibration review is the discipline of shipped outcome vs. predicted outcome, across cycles. It helps a designer read outcomes, investigate drift, and feed validated learning into the next cycle while keeping the final judgment traceable and human-owned.
AI Model & Agent Selection
Choosing the right agent/model for a task, including when general-purpose LLM reasoning beats a specialized tool
Prompt Engineering (general)
Structuring instructions for reliable, scoped agent output, the base-layer skill every stage-specific prompting skill builds on
AI Output Validation
Systematically checking agent output against evidence/spec before accepting it, the practical form of "don't rubber-stamp."
Confidence-Signal Reading
Interpreting an agent's stated certainty level correctly and calibrating scrutiny accordingly
Override & Redirection
Interrupting and correcting an agent mid-task without restarting the whole generation
Ethical AI Judgment
Recognizing manipulative, exclusionary, or biased AI output and vetoing it, independent of stage context
Compliance Literacy (BFSI/Health/Public Sector)
Knowing what data, disclosure, and audit obligations apply to the current engagement's vertical
Complacency Self-Audit
Recognizing one's own automation bias in the moment, the individual-level practice behind the Framework's named systemic risk
Audit-Trail Discipline
Keeping AI-generated artifacts traceable to source and approver as a habit, not an afterthought
Cross-Agent Orchestration
Combining output from multiple agents (e.g., Research + Persona + JTBD Agents) into one coherent artifact rather than treating each in isolation
No matching skills.
Try a broader term or reset the filters.