replay-triage

Prepare for a customer call by synthesizing engagement, sentiment, and feedback from Pendo analytics. Use this skill whenever someone asks about account health, wants to prepare for a customer call, needs an account health report, asks about engagement or usage for a specific customer, or wants a summary of how an account is doing. Also trigger when users mention preparing for a QBR, checking on a customer, or reviewing an account before a meeting — even if they don't say "account health" explicitly.

Generate a shareable health report for one application, summarizing engagement, acquisition, retention, average time, stickiness, frustration signals, and survey scores into a verdict and next steps. Use this skill whenever someone asks how healthy an app is, wants an app health report, wants an overall picture of how an application is doing, or asks about an app's engagement, retention, stickiness, or friction at the application level — even if they don't say "app health" explicitly. This is app-level (a whole application), not account-level — for a single customer account use account-health, and for a single feature use feature-adoption.

Use when planning a product feature, fix, or prioritization task where Pendo usage data, signals, issues, or customer feedback could ground the plan in real product behavior — e.g. "plan a fix for the pricing-page dropoff", "what should I build next", "improve the welcome flow". Skip for pure code refactors, infra, or tooling work.

Analyze feature adoption rates, identify power users vs laggards, and track adoption trends using Pendo analytics. Use this skill whenever someone asks about feature adoption, feature usage, who's using a feature, adoption rates, feature rollout progress, or wants to understand how a specific feature is performing. Also trigger when users mention tracking a feature launch, finding champions or power users of a feature, identifying accounts that haven't adopted a feature, comparing adoption across segments, or analyzing usage trends over time — even if they don't say "adoption" explicitly. If someone asks "who's using X" or "how is feature Y doing", this is the right skill.

Analyze customer feedback using Pendo's feedback tools — cluster themes, extract insights, and surface churn/frustration risks. Use whenever the user asks about feedback trends, top complaints, feature requests, Voice of the Customer, churn risks, or what customers are saying about a topic or account. Triggers on phrases like "feedback report", "what are customers asking for", "any red flags in sentiment", or "what's the feedback looking like". Requires Pendo connector.

Run a guide-review pass over the Guide Signals Pendo has already produced — list the archive, deactivate, and anomaly signals for an app, show each with its evidence and how old it is, then let the user mute a signal or deactivate/archive the guide. Use this skill whenever someone wants to review their guides, clean up stale or declining guides, see which guides Pendo recommends archiving or deactivating, triage guide signals or anomalies, or snooze a signal — even if they don't say "signals" explicitly. Nothing is computed here; the skill reads what the weekly job produced and never changes a guide without an explicit choice.

Creates production-ready Pendo in-app guides as HTML/CSS/JS files, drawing on guide type best practices, org-specific content guidelines, and user-provided context. Trigger this skill whenever a user types /pendo-guides:pendo-guide-creator or uses natural language like "create a guide", "build a Pendo guide", "write a walkthrough", "make an announcement guide", "create an alert", "build a poll", or "write an in-app message". Also trigger when the user describes a guide use case — e.g. "I want to onboard new users", "we're launching a feature and want to notify users in-app", or "I need to collect feedback in the product." Always use this skill proactively when any Pendo guide creation is requested, even if the request is casual or incomplete — the skill handles gathering missing context from the user.

Answer questions about how Orchestrate journeys and emails are doing, read-only. Use when someone asks whether a journey is set up or working as intended, why fewer people got an email than expected or who was missed, whether a journey changed what people did, what Orchestrate achieved over a period (including a QBR, renewal or exec slide), or whether people are getting too many messages. Not for creating or editing journeys (use orchestrate-journeys).

Scan a codebase for deepening opportunities, present them as a visual HTML report, then grill through whichever one you pick.

Diagnosis loop for hard bugs and performance regressions. Use when the user says "diagnose"/"debug this", or reports something broken/throwing/failing/slow.

Review the changes since a fixed point (commit, branch, tag, or merge-base) along two axes: Standards (does the code follow this repo's documented coding standards?) and Spec (does the code match what the originating issue/spec asked for?). Runs both reviews in parallel sub-agents and reports them side by side. Use when the user wants to review a branch, a PR, work-in-progress changes, or asks to "review since X".

Investigate a question against high-trust primary sources and capture the findings as a Markdown file in the repo. Use when the user wants a topic researched, docs or API facts gathered, or reading legwork delegated to a background agent.

This skill should be used when the user wants to "write agent code", "build an agent with ADK", "add a tool", "create a callback", "define an agent", "use state management" — in a project that needs ADK (Agent Development Kit) API patterns and code examples. It provides a quick reference for agent types, tool definitions, orchestration patterns, callbacks, state management, the graph Workflow API, and reference recipes to study. Do NOT use for scaffolding (use google-agents-cli-scaffold) or deployment (use google-agents-cli-deploy).

Set up Claude Code hooks to block dangerous git commands (push, reset --hard, clean, branch -D, etc.) before they execute. Use when user wants to prevent destructive git operations, add git safety hooks, or block git push/reset in Claude Code.

Generate AI music on RunComfy via the `runcomfy` CLI — a smart router across the music-model catalog. Routes to ElevenLabs AI Music Generation (premium 44.1 kHz stereo vocal tracks, 5 s–5 min, $0.0083/s) and ACE Step / ACE Step 1.5 (StepFun-AI open-weights, tag-driven composition, multilingual lyrics, $0.0002–0.0003/s, ~27× cheaper), plus ACE Step audio-inpaint (regenerate a time range inside an existing track) and ACE Step audio-outpaint (extend a track before or after). Picks the right model for the user's actual intent — premium vocal hook, cheap background music library, multilingual pop song, repair a bad chorus, lengthen a 30 s draft into a 2 min cut — and ships each model's documented prompting patterns plus the minimal `runcomfy run` invoke. Triggers on "generate music", "make a song", "AI music", "background music", "instrumental track", "soundtrack", "jingle", "theme music", "royalty-free music", "compose", "music with lyrics", "extend music", "fix this song", "inpaint music", or any explicit ask to generate or edit music.

创建 lark-cli 的自定义 Skill。当用户需要把飞书 API 操作封装成可复用的 Skill（包装原子 API 或编排多步流程）时使用。