发现 Skills

Route broad or ambiguous Konnect requests to the first owning specialist skill. Use when multiple Konnect surfaces are involved or the right owner is unclear, not when the user already named a specific workflow, toolchain, or non-Konnect task.

typst公式中的微分、偏微分方程编写，latex公式转typst。

Use when you have a spec or requirements for a multi-step task, before touching code

REST API design patterns. Auto-activates when building API endpoints, routes, or controllers.

Use when troubleshooting host service connectivity, debugging data loading issues, investigating progress tracking problems, diagnosing environment or configuration issues, or checking Grafana dashboards.

Use when working on fuzzy membership functions, fuzzy sets, fuzzy-to-neural bridging, FuzzyEngine configuration, or V3 fuzzy set definitions.

Use when diagnosing operation failures, stuck or slow operations, querying Jaeger traces, working with Grafana dashboards, debugging distributed system issues, or investigating worker selection and service communication problems.

Checks progress toward Staff Engineer goals. Use when user says "/staff-progress", asks about career progress, wants Staff Engineer tracking, or needs to review staff signals.

Use when working on Claude Code plugins (creating, modifying, testing, releasing, or maintaining) — provides streamlined workflows, patterns, and examples for the complete plugin lifecycle

Guides Google Cloud Storage architecture, bucket/object governance, upload/download patterns, signed URL access, lifecycle management, and security controls. Use when: building file and object workflows on GCP, integrating backend services with Cloud Storage, or hardening storage operations for production.

Use before a release when the visual spec has drifted from the live product, or when design-token-audit passed but the UI still looks off. Compares a browser/device screenshot against the design source (Stitch MCP export, Figma image, or inline DESIGN.md mockup) across five categories: colors, typography, spacing, layout, interactions. Produces a weighted fidelity score and flags specific mismatches — token-audit checks source code, this skill checks the rendered output. Reach for it on phrases like "does it match the design", "design fidelity", "visual regression", "compare to mockup", "design QA", or when a designer reports "this doesn't look right".

Use for read-only health checks of a project where you want to know 'what's the real state right now' WITHOUT fixing anything along the way. Exercises every discoverable feature through the real UI, captures evidence, and classifies each finding by severity (CRITICAL/HIGH/MEDIUM/LOW/INFO). Produces e2e-evidence/audit-report.md with prioritized recommendations but doesn't apply fixes — that's for follow-up work. Reach for it on phrases like 'audit the app', 'pre-release audit', 'compliance check', 'health report', 'what's the current state', or before a major refactor when you want a baseline that documents every existing issue.

Use for web feature validation via Playwright MCP — real browser interactions, cross-browser support (Chromium/Firefox/WebKit), screenshot + DOM snapshot evidence capture, form testing, responsive layouts, console/network error detection. Reach for it when you've picked Playwright as your browser tool (vs Chrome DevTools MCP), or on phrases like 'Playwright validation', 'validate with Playwright', 'browser feature test', 'cross-browser check', 'take DOM snapshot', 'browser automation for validation'. For performance-focused Chrome inspection use chrome-devtools; for the overall web-validation flow use web-validation.

Browser automation and testing using Chrome DevTools MCP (debugging, performance, network inspection) and Browser-Use MCP (human-like UI interaction, form filling, E2E flows). Use when the user needs to test web apps, debug browser issues, analyze performance, fill forms, run E2E user flows, or inspect network/console activity.

name: gpd-submission-health description: Preflight Google Play releases, validate edits, and verify listing completeness with gpd CLI. Use when shipping to production, troubleshooting a failed Play...

Validate MCP (Model Context Protocol) server tools end-to-end after code changes. Use this skill whenever the user is building, modifying, or deploying an MCP server and needs to verify that all registered tools actually work — not just that the server starts. Triggers on mentions of "MCP server", "MCP tools", "tool validation", "tools/list", "tools/call", or any task where Docker rebuild + health check is followed by a need to confirm tool behavior. Also triggers when the user wants an automated feedback loop that detects tool failures, fixes code, redeploys, and re-tests until green before raising a PR. Do NOT use for general API testing, unit tests, or non-MCP servers.

Generates weekly release cycle reports for the team. - preview: What's shipping (run Sunday after release train) - retro: What happened (run Tuesday after prod stabilizes)

Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with web pages, fill forms, take screenshots, or extract information from web pages.

Create Home Assistant entity platforms (sensor, binary_sensor, switch, light, cover, climate, etc.). Use when the user asks to create entities, add a sensor platform, implement a switch, or work with any Home Assistant entity type. Also triggers on mentions of "SensorEntity", "SwitchEntity", "entity platform", "native_value", "device_class", or specific platform names.

Generate and explain Home Assistant service calls for controlling entities (lights, switches, climate, media, covers, etc.) in both Python and YAML. Use when the user asks about calling services, controlling devices, "hass.services.async_call", entity actions, "turn_on", "turn_off", "set_temperature", or needs to invoke HA services from Python code or automations.

Best practices for Home Assistant automations, helpers, scripts, and device controls. TRIGGER THIS SKILL WHEN: - Creating or editing Home Assistant automations, scripts, or scenes - Choosing between template sensors and built-in helpers - Writing or restructuring triggers, conditions, or automation modes - Setting up Zigbee button/remote automations (ZHA or Zigbee2MQTT) - Renaming entities or migrating device_id references to entity_id SYMPTOMS THAT TRIGGER THIS SKILL: - Agent uses Jinja2 templates where native conditions, triggers, or helpers exist - Agent uses device_id instead of entity_id in triggers/actions - Agent modifies entity IDs or config objects without checking all consumers - Agent chooses the wrong automation mode (e.g., using single for motion-activated lights)

Generates Jasmine tests in JavaScript. BDD-style framework with spies and async support. Use when user mentions "Jasmine", "jasmine.createSpy", "toHaveBeenCalled". Triggers on: "Jasmine", "jasmine test", "createSpy", "Jasmine spec".

View Claude Code session logs with colors, filtering, and real-time following

Use just for command running and task automation. Prefer Justfiles over Makefiles. Keep recipes simple - delegate complex logic to scripts rather than embedding it in recipes.