Skip to content

tough-tongue/toughtongue-skills / skills/ttai-agent

ttai-agent

Verified

Creates, 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/entities/account-and-access.md

View on GitHub
# Account and access

Who you are, which workspace you act in, what your plan allows, and who can see
and run a Scenario. MCP reads the workspace with `ttai:get_workspace_info`; it
does not provision users, organizations, members, quotas, or plans.

## Contents

- User and workspaces
- Plan, balance, and entitlements
- Scenario visibility
- Scenario Access Tokens (SAT)
- Agent checklist

## User and workspaces

The human signed in through OAuth owns a **personal** workspace and may belong
to **organizations**. Read both with `ttai:get_workspace_info` (default
sections `me`, `organizations`, `plan`, `balance`); never infer a user ID,
email, or plan from successful login.

- **Personal** — omit `org_id`.
- **Organization** — pass an opaque `organizations[].id` from
  `ttai:get_workspace_info` as `org_id`.

Notes:

- Scenarios, sessions, calls, and bots created with `org_id` belong to that
  organization. No MCP tool moves resources between workspaces.
- `organizations` lists each membership's `id`, `name`, `slug`, the caller's
  `role`, and `quota_enabled`. Never send the slug as `org_id`.
- Roles: `view` < `edit` (Admin in the app) < `owner`; `no_access` = end users.
- `members` (org only, `query` filters name/email) lists people, roles, and
  `quota_available`.
- Inviting members, changing roles, creating organizations, and plan changes
  happen in the web app at
  [app.toughtongueai.com](https://app.toughtongueai.com).
- A `403` means the caller lacks access to that organization or resource: switch
  workspace or have the user fix membership in the app.

## Plan, balance, and entitlements

- **Minutes balance** — `get_workspace_info` section `balance`. Personal wallet,
  or the org wallet plus the caller's quota when quotas are on.
- **Effective plan and limits** — section `plan`. Tier identity/status, limits;
  no price or billing history.
- **Usage** — section `usage` (`start_date` / `end_date`); org editors also get
  a daily series and per-member totals.
- **Quotas** — on: an org session needs minutes in the org wallet AND the
  member's quota. Off: org wallet only. Owners toggle quotas and grant quota
  minutes in the web app; no MCP tool changes them.

Notes:

- Query balance only when the operation may consume it; it is informational.
- Plans gate some features (for example Scenario count, Ocean models, multimodal
  analysis). Check `get_workspace_info` section `plan` or let the server answer; never
  guess a tier from a balance or a Scenario.

## Scenario visibility

- `is_public: true` (default) — anyone with the run link can start a session.
  Phone calls and meeting bots run it with no extra setting.
- `is_public: false` — needs a SAT or an in-app share to open. Phone calls
  still run it (the caller needs edit access). A meeting bot runs it only when
  sent with `session_credentials: "bot_uc"` (the bot runs as the user who sends
  it); without that the server rejects the bot.
- `passcode` — extra gate before start.
- `starts_at` / `ends_at` — availability window.
- `access` — `is_one_time_access`, `max_access_count`.
- `analysis_access` — `default` (run page only), `always` (everywhere), `never`
  (owner share links only).

Run link: `https://app.toughtongueai.com/run/<scenario_id>`. Embeds and iframes
use the same Scenario ID and visibility rules.

**Choose public by default.** Send `is_public: true` on create, including for
drafts, pre-reads, and test Scenarios: a public run link is the simplest share,
and every meeting bot for a private Scenario needs an extra setting. Make a
Scenario private only when the user asks to limit who can open it (a few named
people, a paid cohort). Phone calls keep working; every meeting bot for it is
sent with `session_credentials: "bot_uc"`. Share a private Scenario with a
SAT.

## Scenario Access Tokens (SAT)

Short-lived access tokens, minted by the server, for a private Scenario or an
embed. `valid_for_hours` is 1–168 (default 1). The response carries `type`,
`token`, `expires_at`, and `url` (an iframe embed).

`ttai:create_token` takes one `request` with `type: "scenario_access"`
(`scenario_id`, `valid_for_hours`, optional `email`) and decides who runs the
Scenario with the token. (`type: "filebase_view"` instead links an app
FileBase; see the MCP tools reference.)

- No `email`: an anonymous runner. With `email`: a specific runner (see below).
- `self_billed: true` (no `email`): the caller runs it; usage bills to the
  caller.

Notes:

- Confirm the audience and duration immediately before minting.
- An email-targeted SAT is a consequential write: in an organization it creates
  or reuses a dormant organization user (the email must follow the
  organization's managed-address format the server requires); in a reseller
  context it creates or reuses a real user. Confirm the intended identity.
- Return the link to the user; never store tokens in long-lived notes.

## Agent checklist

1. Resolve workspace (`ttai:get_workspace_info` → `organizations[].id` as
   `org_id`, or personal).
2. Ambiguous team vs personal before a write → ask.
3. Before sharing a private Scenario, mint a SAT or confirm `is_public`.
   A meeting bot for a private Scenario is sent with
   `session_credentials: "bot_uc"`; a phone call runs it as is.
4. Never invent organization IDs, tokens, or plan facts.

## Key Files

- [resources.md](resources.md) — what lives in a workspace
- [scenario-engine.md](scenario-engine.md) — run channels
- [../mcp/tools.md](../mcp/tools.md) — tool conventions
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: 472kUpdated Claude CodeCodexVerifiedInstall
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: 323kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 323kUpdated Claude CodeCodexVerifiedInstall
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 availableStars: 323kUpdated Claude CodeCodexVerifiedInstall