Use for the AI Framing phase (AI/ML products ONLY) — ML problem framing, input/output schemas, evaluation metrics with thresholds, and data strategy. Dispatch after Deep Market Research is approved and before the PRFAQ, when the product involves ML models, predictions, generative AI, or automated decisions.
Use for the Deep Market Research phase — 6 parallel research dimensions, competitor/persona/tech-radar analysis, source validation, and the MarketResearch.html brief. Dispatch when starting product development or when the user asks for market research on a product concept.
Use at the START of the Prototype phase to create the shared [product-slug].css and DesignSystem.html visual reference, and to extract the Design Token Contract — BEFORE any Screen.html is built. Dispatch once, before screen-builders.
Use for the PRD phase — EARS-syntax requirements, personas with dashboard widgets, success metrics, AWS-native technical architecture WITH an inline SVG diagram, and current-year technology research. Dispatch after the PRFAQ is approved.
Use for the PRFAQ phase — Amazon Working Backwards press release + skeptical FAQ, grounded in the market research brief. Dispatch after research is approved and the user wants the PRFAQ.
Use to run a review lens over completed artifacts — accessibility (WCAG AA), full-interactivity/link audit, risk analysis, or competitive response. Dispatch when the user asks to audit, stress-test, or review the prototype or product, or invokes a corresponding slash command.
Use to build ONE prototype screen file (Screen.html) from a screen-builder contract during the Prototype phase. The Orchestrator dispatches one per screen, in parallel. Each invocation builds exactly one screen, links the shared CSS, pastes the sidebar shell verbatim, wires Content Link Map hrefs, and adds the…
Runs after a tool call finishes for Write and Edit tool calls, executing s.js with --check and --noout. From aws-samples/sample-kiro-cli-prompts-for-product-teams.
Use when framing an AI/ML product as a machine learning problem — problem type, input/output schemas, evaluation metrics, data strategy. Triggers on "AI framing", "ML problem framing", "frame the model". AI/ML products only; runs between Deep Research and PRFAQ.
Use when writing a PRD — EARS requirements, personas, success metrics, AWS-native architecture with an inline SVG diagram, and current-year technology research. Triggers on "PRD", "product requirements", "write requirements", "EARS".
Use when writing an Amazon-style PRFAQ (Working Backwards press release + skeptical FAQ) for a product, grounded in market research. Triggers on "PRFAQ", "press release FAQ", "working backwards".
Use when building the interactive HTML prototype — shared CSS, design system, screen manifest, per-screen files, navigation hub. Triggers on "build the prototype", "clickable prototype", "prototype screens".
Use when conducting deep market research for a product — competitors, market sizing (TAM/SAM/SOM), customer personas, technology radar, and a sourced MarketResearch brief. Triggers on "market research", "competitive analysis", "research this product idea", or the start of the product-development workflow.
Use to rebuild the ScreenIndex navigation hub from the Screen.html files actually present in documents/ — after adding, removing, or renaming prototype screens. Triggers on "regenerate the screen index", "rebuild the index", "the index is out of date", "update the navigation hub".
Convert PDFs to synthetic HTML/Markdown via Amazon Bedrock, audit documents for PII with Amazon Comprehend, and generate synthetic tabular datasets (CSV/JSON) from a schema, a preset, a natural-language business description, or a reference document. Use when the user asks to synthesize a document, redact PII, scan a…
LLM grader for behavioral assertions on pocsynth skill evals. Reads a transcript and the eval's expected assertions, emits grading.json per the skill-creator convention.
Generate synthetic tabular datasets (CSV/JSON) for demos, tests, and dashboards — from a bundled business preset, a natural-language description, or a reference document. Validate generated data against its schema. Use when the user asks for fake/test/sample/demo/dummy data, a synthetic dataset, "rows of…
Data Wiki turns data into portable knowledge bundles following Open Knowledge Format (OKF) and serves them to AI agents over the Model Context Protocol (MCP).
Connects Claude Code to the OKF Data wiki consumption MCP server (deployed on Bedrock AgentCore) and provides a skill to fetch/refresh its OAuth2 access token.
23 4d agoA
tokens not measured
MIT-0
At most 3 mods per repository are shown here — the rest are on their repository pages: