plugins-nixtla v1.9.0 — Claude Code plugins + 30 skills for Nixtla time-series forecasting (One-Pager + Operator Audit)

0-plugins-nixtla-one-pager-and-operator-audit.md

plugins-nixtla v1.10.0 — One-Pager + Operator Audit

Claude Code plugins and AI skills for time-series forecasting with Nixtla

Version: 1.10.0 | Plugins: 19 | Skills: 30 | License: MIT

Links: GitHub · Nixtla SDK · TimeGPT

---

One-Pager

The Problem

Time-series forecasting practitioners juggle multiple tools: statsforecast for statistical baselines, TimeGPT for zero-shot deep learning, Airflow for orchestration, Snowflake/BigQuery for data warehousing. Each tool has its own API, learning curve, and integration burden.

The Solution

19 Claude Code plugins that bring forecasting workflows directly into your AI-powered development environment. Each plugin is independently shipped with its own MCP server, tests, and CI workflow.

Tier	Count	Examples
Production v1.0	12	baseline-lab, bigquery-forecaster, snowflake-adapter, dbt-package
Proof of Concept v1.0-poc	2	defi-sentinel, anomaly-streaming-monitor
WIP Scaffold v0.1.0-wip	6	sales-demo-builder, support-deflector, docs-qa-generator

Who / What / Where / When / Why

Question	Answer
Who	Forecasting practitioners, analytics engineers, ML platform owners
What	Claude Code plugins + AI skills for statsforecast, TimeGPT, data warehouses
Where	Claude Code CLI, VS Code, JetBrains IDEs
When	When building, debugging, or explaining time-series forecasting workflows
Why	Reduce context-switching; get expert forecasting assistance in your IDE

Stack

• Python: 3.10+ (Nixtla SDK requirement)
• Core: statsforecast 1.7+, nixtla SDK 0.7.3+
• MCP: Python mcp package, TypeScript @modelcontextprotocol/sdk
• Testing: pytest, Plugin Validator v7.0
• Data: BigQuery, Snowflake, dbt, M4/M5 benchmarks

Key Differentiators

1. Honest Labeling — Every plugin declares its fidelity (v1.0, v1.0-poc, v0.1.0-wip) with tests enforcing the labels
2. Production Hardening — SQL injection mitigation, retry with backoff, lazy imports for fast load
3. One Plugin Per Job — Not a monolith; each plugin ships independently with its own CI gate
4. Offline Baselines — baseline-lab works with zero API keys using statsforecast

---

Operator-Grade System Analysis

Executive Summary

Metric	Value
Version	1.10.0
Total Plugins	19
Production (v1.0)	12
PoC (v1.0-poc)	2
WIP (v0.1.0-wip)	6
Skills	30
CI Workflows	28
Python Version	3.10+

Directory Structure

000-docs/           # 411 markdown files (AAR, specs, standards)
003-skills/         # 30 Claude skills in .claude/skills/nixtla-*/
005-plugins/        # 19 plugin implementations
006-packages/       # nixtla-claude-skills-installer
007-tests/          # E2E/integration tests
tests/              # pytest suite (DEFAULT target)

Plugin Inventory

Phase 1 (Revenue Drivers) — All v1.0:

• nixtla-roi-calculator
• nixtla-forecast-explainer
• nixtla-vs-statsforecast-benchmark
• nixtla-cost-optimizer
• nixtla-migration-assistant
• nixtla-airflow-operator
• nixtla-changelog-automation

Phase 2 (Production Hardening) — All v1.0:

• nixtla-baseline-lab (v1.5.0)
• nixtla-search-to-slack
• nixtla-snowflake-adapter
• nixtla-bigquery-forecaster
• nixtla-dbt-package

Phase 3 (Honest PoC) — v1.0-poc:

• nixtla-defi-sentinel
• nixtla-anomaly-streaming-monitor

Phase 4/5 (WIP Scaffolds) — v0.1.0-wip:

• nixtla-sales-demo-builder
• nixtla-forecast-workflow-templates
• nixtla-forecast-audit-report
• nixtla-support-deflector
• nixtla-docs-qa-generator
• nixtla-embedded-forecast-widget

Operational Commands

# Quick start
./004-scripts/setup-dev-environment.sh
source venv/bin/activate

# Testing
pytest -v --tb=short -m "not integration"

# Validation
python 004-scripts/validate_skills_v2.py --fail-on-warn
bash 004-scripts/validate-all-plugins.sh .

# Formatting
black . && isort .

Security Posture

• No hardcoded secrets in codebase
• API keys via environment variables (.env)
• SQL injection mitigation in bigquery-forecaster (identifier validation)
• No .env files tracked in git

CI/CD

• 28 GitHub Actions workflows (1 per plugin + repo-wide)
• Per-plugin unit tests with pytest -o addopts=""
• Plugin Validator v7.0 marketplace tier gate
• Gemini Code Assist for PR reviews

---

Changelog (v1.10.0)

Release Highlights

Plugin Ecosystem Maturation — 12 plugins ship v1.0, 6 WIP scaffolds added.

Added

• Phase 1 plugins (7): roi-calculator, forecast-explainer, vs-benchmark, cost-optimizer, migration-assistant, airflow-operator, changelog-automation
• Phase 2 plugins (5): baseline-lab hardening, search-to-slack, snowflake-adapter, bigquery-forecaster, dbt-package
• Phase 3 PoC plugins (2): defi-sentinel, anomaly-streaming-monitor
• Phase 4/5 WIP scaffolds (6): sales-demo-builder, forecast-workflow-templates, forecast-audit-report, support-deflector, docs-qa-generator, embedded-forecast-widget
• Honest labeling pattern with dedicated test coverage
• Production hardening patterns (sql_validation, retry, lazy imports)

Fixed

• search-to-slack: Restored 13 of 17 skipped tests
• CI: Vendored validator accepts 'compatibility' field

Technical Details

• Commits since v1.9.0: 21
• Files changed: 180
• Lines added: +14,319

---

Last updated: 2026-05-03 | Generated with /release skill