Owl-Listener/ai-design-skills

harm-anticipation

Proactively identifying failure modes, misuse, and unintended consequences.

state-management

Managing shared context, memory, and state across multiple agents.

multimodal-orchestration

Coordinating text, image, voice, and tool-use modalities in a single interaction.

trust-calibration

Helping users form warranted trust in the AI — neither overtrust nor undertrust — through deliberate confidence and source signalling.

frustration-detection

Reading user emotional state from text signals — caps, punctuation density, repetition, latency — and adapting before the user disengages.

bias-detection-design

Designing review workflows to surface and mitigate bias in AI outputs.

consent-and-agency

Designing for informed user consent, opt-out, and human override.

escalation-design

When and how AI should escalate to humans, refuse, or ask for clarification.

guardrail-design

Defining behavioral boundaries — what the AI should and shouldn't do.

transparency-patterns

Showing users what the AI knows, doesn't know, and how confident it is.

value-specification

Translating organisational values and user expectations into system constraints.

agent-role-design

Defining what each agent does, knows, and owns in a multi-agent system.

failure-recovery

What happens when an agent fails — retry, fallback, escalate, or graceful degradation.

handoff-protocols

Designing smooth transitions between agents and between AI and humans.

human-in-the-loop

Designing intervention points where humans review, approve, or redirect agent work.

observability-design

Making multi-agent workflows visible and debuggable for designers and developers.

task-decomposition

Breaking complex user goals into subtasks that agents can handle.

comparative-evaluation

A/B testing, side-by-side comparison, and preference ranking for AI outputs.

failure-taxonomy

Classifying AI failures — hallucination, refusal, irrelevance, tone mismatch, latency.

heuristic-evaluation-ai

Adapting Nielsen's heuristics and new AI-specific heuristics for AI interfaces.

longitudinal-measurement

Tracking AI product quality over time — drift, degradation, and improvement.

output-quality-rubrics

Defining what "good" looks like for AI outputs — accuracy, relevance, helpfulness.

task-success-metrics

Measuring whether the AI actually helped users accomplish their goals.

user-satisfaction-signals

Interpreting implicit and explicit feedback — edits, regenerations, abandonment.

context-window-design

Designing around token limits, memory, and conversation persistence.

conversation-patterns

Turn-taking, repair sequences, grounding, and dialogue structure for human-AI interaction.

feedback-loops

User correction, thumbs up/down, inline editing, and reinforcement signals.

generative-ui

Designing interfaces where AI generates UI components dynamically.

mixed-initiative-flow

When the AI leads vs. when the user leads, and how to hand off control.

progressive-disclosure

Revealing AI capability gradually to match user mental models.

chain-of-thought-design

Designing reasoning chains that produce better outputs.

constraint-specification

Defining output format, length, tone, and content boundaries within prompts.

context-engineering

Designing what information goes into the context window and in what order.

few-shot-patterns

Crafting examples that steer AI behavior effectively.

prompt-versioning

Managing prompt iterations, testing changes, and tracking what works.

system-prompt-structure

Anatomy of effective system prompts — role, context, constraints, format.

template-design

Creating reusable, parameterised prompt templates for consistent outputs.

behavioral-consistency

Ensuring the AI behaves predictably across sessions, edge cases, and modalities.

cultural-adaptation

Adapting AI behavior for different cultural contexts, languages, and norms.

domain-voice

Tailoring AI behavior for specific professional domains.