tough-tongue/toughtongue-skills / skills/ttai-agent
ttai-agent
VerifiedCreates, edits, and refines Tough Tongue AI Scenarios, the definitions behind voice agents (sales calls, lead qualification, demos, screening, support, booking) and practice roleplays (negotiations, interviews, coaching). Writes ai_instructions, ai_model_config, strategy, tools_config, and rubrik through ttai:create_scenario and ttai:update_scenario. Also resolves the personal or organization workspace, checks plan and balance, places or cancels SIP calls, schedules or cancels meeting bots, mints private Scenario links, grounds a Scenario in the user's documents with knowledge FileBases, and troubleshoots ttai MCP sign-in. Use when the user wants a voice agent or practice Scenario created or improved ("create a voice agent", "the agent ends calls too early", "make it sound more natural"), a call placed, a meeting bot sent, or a Tough Tongue AI account question answered. Not for reports across many sessions (ttai-session-analyst) or browser demo steps (ttai-browser-demo-builder).
references/recipes/cold-call.md
View on GitHub# Cold Call Scenario Patterns
Rules for scenarios where the **AI is the outbound caller** — calling a lead
whose data is (at least partially) known. The user plays the lead.
## Contents
- Voice pipeline (cascade default; half-cascade for cloned TTS + barge-in)
- Opening pattern (two beats, directive in FLOW)
- Lead context handling (template variables, missing-data rules)
- Discovery / data collection
- Objection handling and call outcomes
- ai_instructions and user_instructions structure
- Rubric structure (evaluates the LEAD)
- Technical config, anti-patterns, quality checklist
Key facts:
- **Recommended stamp:** Landmass `cascade` + Cartesia (was `cascade-01` in
older docs). Load [cascade-tts.md](cascade-tts.md). See
[model-selection.md](../scenario/model-selection.md). Ocean realtime is a
fallback only when Cascade is unavailable for the account — never Galaxy.
- Three call sub-types: true cold, warm lead, follow-up — each has different
opening and termination risk
- Two-beat opening: Identity + Reason + Permission → Response branching (4
paths)
- Discovery: collect N data points, ONE question at a time
- Exactly 3 outcomes: CONVERTED / NURTURE / DEAD LEAD
- Rubric evaluates the **lead** (the human), NOT the AI caller
- `skip_auto_start: false` always — AI speaks first
- `end_session` tool required; add a `conductor` timeout (300–450s typical)
---
## Call Sub-Types
Specify the sub-type in `ai_instructions` — it determines the opening pattern
and the lead's expected warmth level.
| Sub-type | Opening | Trust |
| ------------- | ----------------------------------- | ----------------------- |
| **True Cold** | Brief ID + topic + 30s ask | Lowest — high hang-up |
| **Warm Lead** | Name + company + prior touchpoint | Medium — prior intent |
| **Follow-Up** | Name + prior event/role + follow-up | Highest — prior contact |
Examples (directive form in FLOW — not quoted speech):
- True Cold: Hi, is this {{ lead_name }}? Brief call about [topic]. 30 seconds?
- Warm Lead: Hi {{ lead_name }}! [Agent] from [Company]. You recently
[touchpoint]. Still exploring?
- Follow-Up: Hi {{ lead_name }}! [Agent] — [role] at [event] on [date].
Following up.
Default to **Warm Lead** when the sub-type is unclear from the brief.
---
## Voice Pipeline Selection
**Default:** Landmass `cascade` + Cartesia (+ Deepgram STT + Flash Lite LLM).
Include the three cascade-tts blocks (OUTPUT RULES, STT errors, NATURAL SPEECH)
from [cascade-tts.md](cascade-tts.md).
**Fallback:** Ocean `medium-stable` only if Cascade / Cartesia is unavailable
for the account. Do **not** stamp Galaxy for outbound cold calls.
**Half-cascade** (Landmass realtime + external TTS): when voice cloning matters
but you want realtime listening — no full Cascade speech blocks.
**Long, consultative calls.** Cascade stays the default: it gives the most
natural cold-call voice and most teams run it. When the call is a long
conversation with many back-and-forths (discovery beyond a handful of
questions, consultative selling), ask the user whether to try a voice-to-voice
stamp (Landmass `gemini-stable` or Galaxy) instead. Some teams saw better
qualification that way on long calls; it is not consistent, so it is their
call.
---
## Opening Pattern
**Two-beat structure** — never collapse into one turn.
### Beat 1: Identity + Reason + Permission
Max 3 short sentences. State who you are, why you're calling, then STOP.
- Open with the greeting and the name check — suggested text: Hi [first_name]
(when name available) — nothing before it.
- Keep it ultra-short — the shorter the better for TTS delivery.
- **NEVER quote Beat 1.** Quotes make the agent deliver it robotically, and if
interrupted it restarts the quoted block. Write a directive in FLOW. There is
no `strategy.welcome_instructions` field.
```
# WRONG — quoted text causes robotic delivery + restart-on-interrupt:
## FLOW
Say: "Hi Priya! This is Sarah from NovaBridge. Calling about..."
# RIGHT — directive form, agent delivers naturally:
## FLOW
Start with: Hi {{ lead_name }}! This is [Agent] from [Company]. Calling
about [topic]. Ask if they have 30 seconds. Then STOP and wait.
```
### Beat 2: Response Branching (its own turn — never combined with Beat 1)
Script all four branches:
- **YES / WARM** — Great! Just a few quick questions — two minutes. →
discovery.
- **NO / NOT INTERESTED** — No worries! Thanks for your time. Have a great
day! → follow the `end_session` turn rule.
- **BUSY** — Totally get it — two minutes, or I can call back. When works?
- **HOSTILE** — My apologies for the interruption. Have a great day! → follow
the `end_session` turn rule.
Example phrases anywhere in this recipe show tone and content. In
`ai_instructions` label them `tone only` and never ask the agent to repeat them
verbatim; it should rephrase every time. The only `word for word` lines on a
cold call are compliance disclosures and a client's tested script
([ai-instructions.md](../scenario/ai-instructions.md#spoken-lines)).
---
## Lead Context Handling
### Template Variables
```
## CONTEXT
- **Lead name**: {{ lead_name }}
- **Company**: {{ company_name }}
- **Interest area**: {{ interest_area }}
```
### Missing Context Rules (highest priority)
- NEVER speak raw placeholders ("{{ lead_name }}") or filler words ("unknown",
"N/A")
- Name blank → Hi, am I speaking with [role]?
- Name present → treat as guess: Hi, is this [name]?
- Address/context partial → reference what you have, ask to confirm
- Address/context blank → "a [product/property] I have on record under your
name"
- Person corrects data → accept immediately, never re-assert stale info
---
## Discovery / Data Collection
- ONE question at a time — never stack.
- Acknowledge each answer briefly + one short contextual remark → next question.
- NEVER echo back their specific words (STT may have garbled them).
- If unclear, ask to repeat — max twice, then skip.
- Never correct or challenge answers.
- If they decline: No problem at all — skip and continue.
### Progressive disclosure responses
Each data point gets a brief, contextual acknowledgment before the next
question (tone only):
- BAD — flat and robotic: Got it. Next question: what's your budget?
- GOOD — brief warmth + natural flow: Got it — at that scale the right tools
can really make a difference. Quick follow-up: what are you currently using
for that?
Typical data points (5–6 per scenario): context/size, current solution, biggest
pain point, timeline/urgency, budget or decision authority, preferred next step.
---
## Objection Handling (Caller Side)
Core pattern: **Acknowledge → Reframe → Check In**
- **"Not interested"** — probe: priority issue or timing? Accept a firm no
gracefully.
- **"Already have a solution"** — Not looking to replace — anything not quite
working?
- **"Send me an email"** — Happy to — one quick detail so I send the right
stuff?
- **"Too busy"** — Under two minutes — or when's better this week?
- **"How'd you get my number?"** — state the source honestly. Offer removal if
preferred.
Close when: "not interested" is said twice; a removal request is made; the
person is hostile or abusive; the number is wrong; or the lead already completed
the action. Speak the closing first, then follow the live `end_session`
instructions; do not call it in the same turn as a question or pitch.
---
## Call Outcomes (Exactly 3)
| Outcome | Condition |
| ------------- | ------------------------------------------ |
| **CONVERTED** | Qualified + committed next step |
| **NURTURE** | Partial data, door open, not ready |
| **DEAD LEAD** | Not interested, wrong number, DNC, hostile |
Close each outcome, then follow the closing turn rule:
- **CONVERTED** — confirm date/time/action + warmth.
- **NURTURE** — agree on a follow-up plan.
- **DEAD LEAD** — one polite goodbye.
---
## Compliance and Sensitivity Rules
Include a compliance section in `ai_instructions` whenever the domain is
regulated or the call topic is sensitive.
### Universal minimums (all cold calls)
```
## COMPLIANCE
- Identify your company name clearly within the first 30 seconds.
- Honor do-not-call requests immediately — never argue or delay.
- Never impersonate a government agency, bank, lender, or legal representative.
- Never make guarantees about outcomes (financial returns, admission, legal results).
- If asked to stop calling: state the removal confirmation, then follow the
closing turn rule.
```
### Domain-specific additions
**Financial / Wealth Management:** Never give investment advice, quote returns,
or suggest specific securities. Identify as a licensed professional if
applicable.
**Real Estate:** Fair Housing applies — never steer based on protected class.
Never speculate on floor prices or invent competing offers.
**Healthcare / Education (minors):** Never share student/patient data. Never
make admission or medical outcome guarantees. Speak with parent/guardian.
**Foreclosure / Hardship:** Lead with empathy. Never pressure decisions under
emotional duress. Never impersonate the lender/servicer. Offer the opt-out
early. Use `disconnectDelaySeconds: 8`.
---
## ai_instructions Structure
```
## PERSONALITY and TONE
- Personality: [e.g. Warm, efficient, consultative]
- Tone: [e.g. Friendly, professional, low-pressure]
- Length: 2–3 sentences per turn.
- Name usage: sparingly — once at opening, maybe once more.
## CONTEXT
- [Variable]: {{ variable_name }}
## CRITICAL — MISSING CONTEXT RULES
[Missing/wrong data handling per Lead Context section above]
## ROLE
You are [Name], [title] at [Company]. [One-line mission.]
## FLOW
### Phase 1: Opening
[Directive Beat 1. Then STOP. Beat 2: all 4 branches.]
### Phase 2: Discovery (N questions)
[ONE question per turn. Q1–QN.]
### Phase 3: Objection Handling
[Acknowledge → reframe → check in]
### Phase 4: Q&A + next step (keep the call OPEN)
### Phase 5: Close
[State the goodbye. Wait for the lead's reply, then call end_session if they
have not substantively re-opened the conversation.]
## TOOLS
Call end_session for CONVERTED / NURTURE / DEAD LEAD only after the lead's
post-goodbye reply.
## HANDLING QUESTIONS
[FAQ guidance for likely product/company questions]
## GUARDRAILS
## THINGS YOU MUST NEVER DO
[5–8 hard prohibitions]
## COMPLIANCE NOTE (if regulated)
## EDGE CASES
[6–10 situations]
## STYLE
Spoken output rules + speech style. Cascade only: STT-error handling.
```
---
## user_instructions Structure
```
You are [role/persona] who
[context: filled a form / visited an open house / etc.].
[1–2 sentences of emotional texture: busy, skeptical, stressed, curious.]
Respond naturally in [language].
[Optional permissions: "feel free to be price-sensitive", "you may not
remember filling the form", "you might be in a hurry"]
```
Keep short and evocative — tell the human how to _play the lead_, not how to
sell.
---
## Rubric Structure
**CRITICAL**: Cold call rubrics evaluate the **LEAD** (the human), not the AI.
It is a lead-screening brief for the sales/ops team — NOT a performance review.
```
# [Company] Cold Call Report
Evaluate the LEAD — not the AI. Context: [who / why].
## Qualification Score (1-10)
- 10: All data points + strong intent signal
- 8-9: Most data, cooperative, clear interest
- 5-7: Partial data, some interest
- 3-4: Minimal engagement, callback needed
- 1-2: Didn't connect / declined / wrong number
## Structured Data (for [Team])
**Call Outcome**: [fully qualified / partial / callback / not interested /
didn't connect]
**Next Step**: [demo / callback / materials sent / none]
[Field 1]: / [Field 2]: / ...
**Additional Notes**: [Anything volunteered beyond structured Qs]
## Follow-Up Recommendation
**Should [team] follow up?**: Yes — priority / Yes — standard / No
**Recommended action**: [One clear next step]
```
---
## Technical Config: recommended default (Cascade)
```json
{
"ai_model_config": {
"provider": "Landmass",
"model": "cascade",
"tts_provider": "cartesia",
"tts_voice_id": "f786b574-daa5-4673-aa0c-cbe3e8534c02",
"llm_provider": "google_vertex",
"llm_model": "gemini-3.1-flash-lite",
"stt_provider": "deepgram"
},
"strategy": {
"skip_auto_start": false,
"system_instructions_template": "minimal",
"silence": {
"silence_threshold": 5000,
"end_session": false,
"force_agent_to_speak": true
},
"conductor": {
"enabled": true,
"messages": [
{
"time_seconds": 350,
"message": "Wrap up — confirm data or next step and say goodbye. Wait for the lead's reply before end_session.",
"end_turn": true
}
]
}
},
"tools_config": {
"tools": {
"end_session": {
"should_register": true,
"add_to_system_prompt": true,
"tool_settings": { "disconnectDelaySeconds": 8 }
}
}
},
"session_analysis": { "is_auto_analysis": true, "is_auto_submit": true },
"is_recording": true
}
```
Cartesia ID = American female from model-selection.md. Load
[cascade-tts.md](cascade-tts.md) into `ai_instructions`.
### Ocean fallback (Cascade unavailable)
```json
{ "ai_model_config": { "provider": "Ocean", "model": "medium-stable" } }
```
Same strategy / tools / session_analysis as above. No cascade-tts blocks.
---
## Anti-Patterns
| Anti-pattern | Instead |
| ---------------------------------- | --------------------------------------- |
| Beat 1 + Beat 2 in one turn | Stop after Beat 1 |
| Discovery Q on the opening | Wait for a response branch |
| Two questions in one turn | One question per turn |
| Echoing the lead verbatim | Brief ack + context |
| Quoted or fenced example lines | Plain text, labelled tone only |
| Pushing past "not interested" | Graceful close → `end_session` timing |
| Rubric scoring the AI | Score the LEAD |
| Breaks when a `{{ var }}` is blank | Missing-data fallbacks |
| `skip_auto_start: true` | Always `false` |
| No conductor timeout | Always set |
| Galaxy or casual Ocean stamp | Cascade default; Ocean only as fallback |
---
## Quality Checklist
- [ ] Two-beat FLOW opening; 4 branches; 3 outcomes; one Q per turn
- [ ] Example lines unquoted, labelled tone only / word for word
- [ ] PERSONALITY/TONE, HANDLING QUESTIONS, EDGE CASES, NEVER list
- [ ] Missing-data fallbacks; compliance if regulated
- [ ] `end_session` + `skip_auto_start: false` + talkative silence + conductor
- [ ] Stamp: Landmass `cascade` + real Cartesia ID + cascade-tts blocks (or
documented Ocean fallback)
- [ ] Rubric scores the LEAD, not the AI
## Key Files
- [../scenario/authoring.md](../scenario/authoring.md) — durable design rules
- [../scenario/model-selection.md](../scenario/model-selection.md) — pipeline
stamps
- [../scenario/control.md](../scenario/control.md) — strategy / tools
- [cascade-tts.md](cascade-tts.md) — full Cascade only
- [../scenario/rubric.md](../scenario/rubric.md) — evaluation text
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |