callflows

voice-agent-kit

Build, change and ship ElevenLabs voice agents. Use this skill for anything voice-agent related — a new phone agent for a customer, changing an existing agent's behavior, opener or voice, a test call, or post-call task extraction. Runs a fixed pipeline from briefing to deployed agent.

vac-campaign-goal

Generates the campaign configuration for a voice-agent build — campaign-goal.md (the campaign goal definition, i.e. the campaigns.campaign_goal free-text field) plus call-outcomes.json (the typed campaigns.call_outcomes list). Runs at step 9 of the VAC pipeline (per the SSOT in PIPELINE.md), after vac-task-extraction. Both artifacts are derived from the agent's briefing.md Call Objectives and are what the campaign platform uses at runtime to classify each call's outcome (success / lead) and to feed the goal context into every post-call analysis. Use whenever building a voice agent and the campaign's goal definition + outcomes are needed, or whenever the user asks to create a goal definition, campaign goal, call outcomes, success definition, outcome types, or the campaign platform's campaign-goal/outcome config for an agent — even if they only say "generate the goal" or "outcomes for agent X" without naming the fields. English output artifacts.

vac-company-research

Research the calling/represented company from its website/domain. Dual role - (1) Step 0 auto-intake bootstrap when no human intake is available; (2) the DETAIL source for the agent's knowledge base. The few core facts + call purpose feed the prompt (tier-1); detailed company Q&A feeds the KB (tier-2, via vac-knowledge-base). Scrapes the company website for industry, products, target audience, brand voice, and prospect-level detail.

Code & Development Listed

vac-deploy

Deploy a voice agent directly to ElevenLabs and store artifacts in GitHub. Use after vac-review PASS. Handles voice selection, deployment via ElevenLabs API (EU Residency), and git commit/push.

vac-design

Design the conversation architecture for a voice agent. Use after research to define conversation phases, transitions, dynamic variables, guardrails, behavioral modules, and no-gos. Produces the conversation blueprint that the prompt writer follows.

API & Backend Listed

vac-elevenlabs-api

ElevenLabs Conversational AI Agent API wrapper — full CRUD (Create, Read, Update, Delete) plus Outbound Calls. Direct API access against EU Residency endpoint.

vac-intake

Structured intake and briefing for new voice agent builds. Use when starting a new voice agent project. Collects customer context, use case, target persona, call objectives, available data, and compliance requirements. Detects gaps and classifies the use case.

vac-knowledge-base

Decide what goes in the prompt vs. a RAG knowledge base, then author the KB document for an engaged-decision-maker voice agent. Use after vac-design (when it flags "KB required") and BEFORE vac-prompt, so the prompt writer knows the prompt-vs-KB split. Produces knowledge-base.md (tier-2 substance, RAG-style) + a tier-1 spec for the prompt. CONDITIONAL step — not every agent needs a KB.

vac-prompt

Write the ElevenLabs voice agent prompt and configuration. Use after conversation design to produce the actual prompt (English) and config.json following ElevenLabs best practices. Outbound agents get empty first_message (callee speaks first). Optimized for low latency and natural conversation quality.

vac-research

Research and select conversation frameworks for voice agent builds. Use after intake to determine which sales/conversation frameworks best fit the use case. Draws from curated framework library and researches new frameworks when needed.

Code & Development Listed

vac-review

Review voice agent prompts against a 20-dimension scoring rubric. Use after prompt writing to score quality and identify issues. Iterates up to 3 times, then escalates to a human reviewer (the requester). Minimum score 4.0/5.0 required.

Data & Documents Listed

vac-task-extraction

Generates a task-extraction-prompt.md for a specific voice agent build. This prompt is executed at runtime (after each call) to analyze transcript + call summary and produce CRM tasks for human follow-up. Use at step 8 of the VAC pipeline (see PIPELINE.md), after vac-test. The output file contains ONLY the bare prompt, copy-paste ready; applied defaults and notes are reported in chat, not written into the file. Default output is a prose title+description task; only when the briefing prescribes or clearly implies a structured multi-field task does it also emit a conditional extraction-schema.json (the per-campaign structured-output schema your runner uses to shape the output). Trigger whenever building a voice agent and a post-call CRM task extraction is needed, or whenever the user asks to create a task extraction prompt, post-call processing, CRM task generation, or transcript analysis for a voice agent build.

vac-test

Generate test scenarios and evaluate test call transcripts for voice agents. Use after deployment to assess real-world performance. Feeds findings back into prompt iteration. Requires human test callers.