Skip to content
Back to skills

Lytics Audiences

ASecurity

Lytics CDP: build, advise on, analyze, and manage audience segments. Translates natural language into FilterQL, validates and sizes it, and creates or updates the segment; recommends audience strategy for a business goal or improves an existing segment using ML feature importance, content affinities, and fieldinfo; shows audience composition (demographic breakdowns, top field values, coverage rates, distributions, segment comparisons); and lists, views, creates, updates, deletes, validates, s...

  • 4 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added October 10, 2026
ai-agentsgoexpressapi

Works with

  • api

Security analysis

A100/100

Pro scans all 10 files and shows the line behind each finding

Scanned October 10, 2026

npx -y skills add lytics/agent-skills --skill lytics-audiences --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Lytics Audiences?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Lytics Audiences
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/lytics-lytics-audiences/badge)](https://www.skillsdirectory.com/skills/lytics-lytics-audiences)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: lytics-audiences
description: "Lytics CDP: build, advise on, analyze, and manage audience segments. Translates natural language into FilterQL, validates and sizes it, and creates or updates the segment; recommends audience strategy for a business goal or improves an existing segment using ML feature importance, content affinities, and fieldinfo; shows audience composition (demographic breakdowns, top field values, coverage rates, distributions, segment comparisons); and lists, views, creates, updates, deletes, validates, sizes, and reevaluates segments. Use when the user wants to build, create, or define an audience or segment from a description; write, build, or validate a FilterQL query or filter expression; needs help choosing an audience strategy, advice on segment design, or to improve an existing segment; wants to understand audience composition, view segment demographics, or analyze field coverage for a segment; or wants to list, view, update, delete, validate, or size segments."
license: MIT
---

# Lytics Audiences

Everything about audience segments in Lytics: turning a description into valid FilterQL, coaching toward the right audience for a business goal, describing who is in a segment, and the segment lifecycle (list, get, create, update, delete, validate, size, reevaluate).

## Before you start

- Credentials: `references/auth.md`.
- Request and response conventions (two response shapes, URL-encoding, status codes): `references/api.md`.
- FilterQL syntax and operator/type compatibility: `references/filterql-grammar.md`, `references/field-types.md`.
- Every create, update, delete, and reevaluate goes through `references/confirmation-gate.md`. Reads (list, get, size, validate, fieldinfo) run immediately.

## Gotchas

- **Slug dedupe on create**: there is no 409. A taken `slug_name` is silently renamed to `<slug>_1`, `<slug>_2`, ... and the create still succeeds. Always report `.data.id` and `.data.slug_name` from the response, never the slug you sent, and say so explicitly if they differ. (On update, a taken slug is a 400 `Slug is already used.`)
- **Update can create a different segment**: if the new `segment_ql` omits `FROM`, the table resets to `user`; for a non-user segment the PUT misses the original and **creates a new segment**, and still succeeds. GET first, keep its `FROM <table>`, and check the returned `.data.id` equals the id you PUT to. If not, stop and tell the user a new segment was created.
- **Update can rename the slug**: if `ALIAS` differs from the current slug and no `slug_name` is sent, the slug is renamed, breaking every segment that does `INCLUDE <old_slug>`. Keep the existing `ALIAS <slug>`.
- **fieldinfo needs the `id` hash, not the slug**: `/api/segment/<slug>/fieldinfo` returns HTTP 500. Resolve the id via `GET /v2/segment` first.
- **validate and size take raw FilterQL text** (`Content-Type: text/plain`), not JSON.
- **`fieldsuggest` requires `q`**: `/api/schema/user/fieldsuggest/<field>?q=<query>`.
- Delete is irreversible; say so at the confirmation gate.

## Modes

Read the mode file before acting in that mode.

| Mode | When | File |
|---|---|---|
| build | The user describes a specific audience ("US users who visited 5+ times") or wants a FilterQL expression written or validated | `build.md` |
| advise | The user states a business goal ("drive more purchases") or wants an existing segment improved | `advise.md` |
| snapshot | The user wants to know what a segment looks like: demographics, top values, coverage, distributions, comparisons | `snapshot.md` |
| manage | List, get, ancestors, create from a ready payload, update, delete, validate, size, reevaluate | `manage.md` |

Typical sequence: advise -> build -> snapshot. A goal-oriented build request should start in advise; after any create, offer a snapshot.

## Related skills

- `lytics-schema`: find which fields exist, their types, and value distributions before mapping concepts to fields; add a missing field.
- `lytics-profiles`: look up an individual profile and the segments it belongs to.
- `lytics-content`: set up or curate content topics when affinities are not configured for advise mode.
- `lytics-integrations`: export or activate a segment to a destination once it exists.
- `lytics-flows`: use a segment as a flow entry audience.
- `lytics-account`: copy a segment to another account.

Files in this skill

  • SKILL.md4.3 KB
  • advise.md7 KB
  • build.md6.4 KB
  • manage.md4.2 KB
  • references/api.md2.5 KB
  • references/auth.md4.1 KB
  • references/confirmation-gate.md2.1 KB
  • references/field-types.md4.2 KB
  • references/filterql-grammar.md4.4 KB
  • snapshot.md3.4 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.