diagnose-call

Create, update, test, and attach Synthflow agent actions over MCP — custom HTTP actions, information extractors, transfers, SMS, real-time booking (Cal.com, GoHighLevel), IVR navigation, custom evaluations, and call-summary emails. Use this whenever the user wants to give an agent a new capability, connect it to an API or CRM, extract data from calls, add a transfer or SMS, set up booking, change an action's URL, body, or variables, or fix an action that isn't firing or is failing. Trigger on phrases like "add an action", "call my API during the call", "extract the caller's email", "set up a transfer", "connect booking", "update the webhook URL", or "the action isn't working".

Build a call-volume and outcome report for a Synthflow agent, workspace, or agency — counts, durations, statuses, end-call reasons, call reasons from information extractors, transfer and booking results, and trends over a period. Use this whenever the user wants a daily, weekly, or monthly report, a call-reason breakdown, KPIs for a customer, an operations dashboard dataset, or numbers like "how many calls did we get", "what are people calling about", or "what share of calls were transferred". For a quality audit of individual calls, use the call-review skill instead.

Review the last 100 calls for an agent and surface problems. Use this whenever the user wants a call-quality audit, a QA pass over recent calls, a "what's going wrong with this agent" review, or to find failed, unresolved, dropped, or otherwise problematic calls. Trigger on phrases like "review my calls", "what's wrong with the calls", "QA the last 100 calls", "find bad calls", or "audit recent calls".

Review AI voice-agent prompts, especially Synthflow Single-Prompt, for contradictions, missing context, unsafe tool/action behavior, escalation gaps, compliance risk, and regression risk. Use this when the user asks to review, audit, validate, improve, or compare a prompt before deployment.

Build and run simulated-caller tests against a Synthflow voice agent, then report which scenarios pass or fail with transcript evidence. Use this whenever the user wants to test an agent before or after a change, check that a prompt fix works, reproduce a failure from a real call, build a test suite, or re-run existing simulations. Trigger on phrases like "test the agent", "run the simulations", "regression test", "make sure this didn't break anything", "build a test suite", "reproduce this call", or "re-run the tests".

Make a change to a live Synthflow voice agent safely — record a baseline, save a draft, review the diff, publish only on approval, verify the published state, and roll back if needed. Use this whenever the user wants to edit, update, fix, tweak, or publish an agent's prompt, greeting, voice, transcriber, LLM, actions, or any other setting, or wants to roll an agent back. Trigger on phrases like "update the prompt", "change the voice", "fix the agent", "publish the draft", "push this live", "roll back the agent", or "undo the last change".

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。