v0.4.0 — Powered by altimate-core

Now uses altimate-code check for 26 lint rules, DataFusion validation, policy guardrails, PII detection, and safety scanning. Requires altimate-code >= v0.5.10. 408 tests. See docs/v2-migration.md for upgrade guide.

v0

What's Changed

• feat: use altimate-code check for deterministic SQL analysis by @anandgupta42 in #3

New Contributors

• @anandgupta42 made their first contribution in #3

Full Changelog: v0.3.0...v0

v0.3.0 — Mermaid DAG + Executive Summary

Altimate Code Actions v0.3.0

The PR comment that makes data engineers say "wow."

What's New

Executive One-Line Summary — Everything at a glance:

## ⚠️ Altimate Code — `2 models` modified · `12 downstream` · `2 warnings` · `+$1.60/mo`

Mermaid DAG Blast Radius — Colored dependency graph rendered by GitHub:

graph LR
    classDef modified fill:#ff6b6b,stroke:#333,color:#fff
    classDef downstream fill:#ffd93d,stroke:#333
    classDef exposure fill:#845ef7,stroke:#333,color:#fff
    stg_orders:::modified --> fct_revenue:::downstream
    fct_revenue:::downstream --> exec_dashboard:::exposure

Cost Before/After Table:

Model	Before	After	Delta	Cause
fct_revenue	$0.80/mo	$2.40/mo	+$1.60	Missing partition filter

Critical Issues Auto-Expand — Not hidden behind collapsible sections.

Fix Suggestions Inline — Every issue shows how to fix it.

Comment Length

• Clean PR: ~6 visible lines
• Warnings: ~10 visible lines
• Critical + impact + cost: ~12 visible lines, ~70 expanded

Stats

• 344 tests, 0 failures
• 19 SQL rules across 4 categories
• 1.1 MB bundled dist/index.js

v0.2.0 — Redesigned PR Comments + Interactive Commands

Altimate Code Actions v0.2.0

What's New

Redesigned PR Comment — Compact, scannable, competitive with CodeRabbit/SonarCloud:

• Summary table showing all checks at a glance (~8 visible lines)
• Issues grouped by severity in collapsible sections
• ASCII DAG visualization for pipeline impact
• Cost before/after/delta per model

Inline Review Comments — Critical issues posted directly on diff lines:

• Batched as a single GitHub review (no notification spam)
• Cap at 10 inline comments per PR

Interactive Commands — @altimate review, @altimate impact, @altimate cost, @altimate help

19 SQL Rules (was 14) — 5 new rules with fix suggestions on every rule:

• function_on_indexed_column, not_in_with_nulls, distinct_masking_bad_join, count_for_existence, no_limit_on_delete
• All rules categorized: correctness, performance, style, security

Quick Start

# Zero config — no API key needed
- uses: AltimateAI/altimate-code-actions@v0
  with:
    mode: static

Stats

• 329 tests (was 240), 0 failures
• 1.1 MB bundled dist/index.js
• 19 built-in SQL rules across 4 categories

v0.1.0 — AI-Powered SQL/dbt Code Review (Beta)

Altimate Code Actions v0.1.0 (Beta)

AI-powered SQL & dbt code review for every pull request. Quality gates for data teams.

Quick Start

Zero-config static analysis (no API key needed):

- uses: AltimateAI/altimate-code-actions@v0
  with:
    mode: static

With AI review:

- uses: AltimateAI/altimate-code-actions@v0
  with:
    model: anthropic/claude-haiku-4-5-20251001
  env:
    ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}

Features

• SQL quality analysis — 14 built-in anti-pattern rules
• dbt DAG impact analysis — downstream blast radius on every PR
• Query cost estimation — Snowflake & BigQuery
• PII detection — 15 categories flagged before merge
• Schema breaking change detection
• Multi-dialect — Snowflake, BigQuery, PostgreSQL, Databricks, Redshift
• Interactive mode — @altimate mentions for on-demand analysis
• Configurable via .altimate.yml

Test Coverage

• 240 tests (110 unit + 130 E2E), 0 failures
• 7 real open-source dbt project fixtures
• dbt version matrix: 1.7, 1.8, 1.9
• 5 SQL dialect test suites

What's New in This Release

• model is now optional — static analysis works without any API key
• Comment body truncation prevents GitHub 422 errors on large PRs
• .altimate.yml config file now works from the main action
• Prompt injection mitigation for SQL content
• Fork PR support (results written to job summary)
• Retry logic with exponential backoff on all GitHub API calls
• Credential masking in logs
• Complete dbt CI documentation (profiles.yml, dbt deps, manifest)
• Databricks and Redshift warehouse connection examples
• Monorepo example workflow