Altertable Skills

CI Score Skills Agent Skills License: MIT Skills Python 3.12+

A collection of Agent Skills for Altertable AI agents. Skills are portable, version-controlled packages of instructions, scripts, and references that give agents new capabilities.

What are Agent Skills?

Agent Skills are an open standard for giving AI agents specialized capabilities. Each skill is a self-contained folder with a SKILL.md file containing instructions that agents load on demand. Build once, use across any compatible platform.

Quick Start with /altertable:ask

Don't know which skill to use? Just type:

/altertable:ask <your question>

The /altertable:ask command routes your query to the best skill automatically. Examples:

  • /altertable:ask show me my web traffic -> analyze-web-traffic
  • /altertable:ask what tables do I have? -> explore-data
  • /altertable:ask analyze my signup funnel -> analyze-funnels
  • /altertable:ask what is Altertable? -> understand-platform

Available Skills

Skill Description
analyze-funnels Builds conversion funnels over ordered steps
analyze-insights Explains what an existing insight or chart shows
analyze-web-traffic Analyzes website traffic (pageviews, sessions, referrers, landing pages)
ask Routes user queries to the best-fit Altertable skill
build-segments Compares event metrics across cohorts using filters, breakdowns, and dimensions
configure-tasks Schedules recurring AI tasks over insights and dashboards
create-discoveries Reports a change, anomaly, root cause, recommendation, or warning to users
create-insights Drafts, renders, and saves insights of every type (SQL, semantic, segmentation, funnel, retention)
decide-actions Decision matrices for picking insight types (funnel, retention, semantic, segmentation, SQL), task types, and discovery actions
evaluate-skills Evaluates and authors agent skills against the Agent Skills spec
explore-data Inspects catalogs, schemas, tables, columns, semantic models, measures, and dimensions
forecast-timeseries Runs on-demand statistics over a time series to spot outliers and project values
instrument-product-analytics Adds Altertable product analytics to an application (event tracking, user identification, traits, consent, session reset, aliasing)
manage-discoveries Reviews, approves, and rejects existing discoveries
query-lakehouse Writes, validates, optimizes, and runs DuckDB SQL against the Altertable lakehouse
query-product-events Queries product events and identities with SQL (event counts, properties, user activity, traits)
understand-platform Explains Altertable concepts and architecture
use-memory Stores and recalls agent memories and org knowledge between sessions

Getting Started

git clone https://github.com/altertable-ai/skills.git
cd skills
uv sync
uv run pre-commit install

Validate a skill

uv run skills validate ./skills/skill-name

Score a skill

Every skill is scored by an LLM judge against the Agent Skills spec (threshold: 70/100):

uv run python scripts/score-skills.py ./skills/skill-name --verbose

Run tests

uv run pytest scripts/tests/ -v

Creating a New Skill

cp -r templates/SKILL_TEMPLATE skills/my-new-skill

Each skill follows the Agent Skills Specification:

skills/
  skill-name/
    SKILL.md          # Required: metadata + instructions
    references/       # Optional: detailed documentation
    scripts/          # Optional: executable code
    assets/           # Optional: templates, resources

See CONTRIBUTING.md for full guidelines.

Compatibility

Skills follow the open Agent Skills standard and work across 30+ compatible platforms:

Claude Code | Cursor | VS Code | Gemini CLI | OpenAI Codex | GitHub Copilot | Goose | Roo Code | OpenHands | and more

Resources

License

This project is licensed under the MIT License. See LICENSE for details.