Use this skill when the user wants to generate ideas, challenge a plan, or think through a problem from multiple angles. Triggers include: 'brainstorm', 'what are my options', 'red-team this', 'what could go wrong', 'devil's advocate', 'pre-mortem', 'give me ideas', 'challenge this', 'what am I missing', 'generate alternatives', 'stress-test this plan', or any request to explore possibilities or find weaknesses before committing. Do NOT use for user research synthesis — use ai-maker-user-research for that.
Use this skill when the user wants a visual output that renders as an HTML page — dashboards, status boards, trackers, comparison tables, timelines, or any rich visual deliverable. Triggers include: 'build a dashboard', 'create a canvas', 'make a status board', 'visual summary', 'render this as HTML', 'interactive report', 'one-pager layout', 'card grid', 'progress tracker', or any request where the output should be a rendered HTML artifact rather than plain text. Also use when the user asks for an output they can share or print as a visual document.
Use this skill when a manager needs to understand, review, or evaluate code without writing it themselves. Triggers include: 'explain this code', 'review this PR', 'what does this do', 'is this good code', 'evaluate this architecture', 'what are the risks', 'prepare questions for the engineering team', 'what's the technical debt here', or any request to understand or assess a technical codebase from a manager's perspective. Do NOT use for writing production code or debugging implementations — use technical Workbench skills for that.
Use this skill when the user needs to interpret metrics, understand what data means, build a data visualization, or define what to measure. Triggers include: 'what does this data mean', 'interpret these metrics', 'build a chart', 'create a dashboard', 'show me a trend', 'what should we measure', 'data summary', 'metrics report', 'anomaly in the data', or any request to make sense of numbers for a business decision. Do NOT use for building technical data pipelines or processing large datasets — use ai-workbench-data for that.
Use this skill when the user wants to create an HTML artifact, dashboard, one-pager, or visual report, or when they want feedback on a design's visual quality or brand compliance. Triggers include: 'build a dashboard', 'create an HTML page', 'design a one-pager', 'review this design', 'does this match the brand', 'make it look better', 'add a visual layout', 'canvas output', or any request to produce or review a polished visual deliverable. Do NOT use for technical UI development (React, WinUI 3) or accessibility audits of existing code — use the appropriate technical skill for those.
Use this skill for management communication, sprint operations, and team coordination. Triggers include: 'draft an email', 'write a status update', 'plan this sprint', 'write a meeting agenda', 'capture action items', 'help me communicate this', 'translate this for the team', 'stakeholder update', 'sprint planning', 'retrospective', 'escalation', or any request to run an effective team operation. Do NOT use for sending actual emails or scheduling meetings in M365 — use ai-maker-workiq for that.
Use this skill when the user needs to define testing strategy, set quality standards, evaluate test coverage, or review a QA plan. Triggers include: 'what should we test', 'is our test coverage good enough', 'define acceptance criteria', 'review this QA plan', 'what's the right testing approach', 'how much testing do we need', 'quality gate', or any request to reason about what quality means for a project. Do NOT use for writing actual test code — use ai-workbench-testing for that.
Use this skill when the user wants to research a topic, evaluate a vendor or product, build a competitive landscape, or make an evidence-backed decision. Triggers include: 'research this', 'competitive analysis', 'what does the data say', 'investigate', 'market landscape', 'evidence for', 'look into', 'compare options', 'due diligence', 'strategic analysis', or any request to synthesize multiple sources before making a decision. Also use when preparing a research brief, executive summary, or recommendation memo. Do NOT use for deep technical implementation research (API docs, GitHub repo analysis) — use ai-workbench-researcher for that.