Analysis

Connect your repos. Get results in minutes.

Octokraft detects your languages, selects the right analyzers, and runs everything automatically. No config files, no setup wizards, no YAML to maintain.

Your stack is detected, not configured

When you connect a repository, Octokraft scans for languages and selects the appropriate analyzers. If you already have linter configs like ruff.toml, biome.json, or .rubocop.yml, those are picked up automatically.

  • Python, JavaScript, TypeScript, Go, Rust, Java, Kotlin, C#, Ruby, and PHP supported out of the box.
  • Existing linter configurations are respected. Your team's rules stay intact.
  • Build artifacts and dependency directories are excluded automatically. No need to manually ignore node_modules or target/.
Detected analyzers table showing language, path, and ignore patterns

Five analysis sources, one unified result

Every repository is analyzed by five complementary sources that run in parallel. Results are normalized to a unified severity scale and deduplicated so the same issue is never reported twice.

  • Static analyzers run language-specific linters like Ruff, Biome, Clippy, golangci-lint, and more.
  • AI-powered semantic analysis catches logic errors, missing edge cases, and architectural concerns that linters miss.
  • Code graph analysis maps symbols, dependencies, and cross-file relationships to detect dead code, circular dependencies, and coupling.
  • Convention detection learns the patterns your team follows and flags code that deviates.

Languages and analyzers

Each language gets purpose-built static analysis plus AI review. No plugins to install.

Python

Ruff + Bandit

JavaScript / TypeScript

Biome

Go

golangci-lint

Rust

Cargo Clippy

Java / Kotlin

Checkstyle + PMD

C#

dotnet-format

Ruby

RuboCop

PHP

PHPStan

Analysis runs when it matters

You don't trigger anything manually. Octokraft hooks into your workflow and runs analysis at the right moments automatically.

  • Repo connected: full onboarding runs automatically. Clone, detect, analyze, compute health score. Most repos finish in 1 to 3 minutes.
  • PR opened or updated: classification, risk signals, merge readiness, and findings posted directly to the PR.
  • PR merged: incremental health assessment recalculates scores for changed files.
  • Code pushed to default branch: code graph and conventions updated to reflect the current state.

Connect your repos and see for yourself

No config files. No setup wizard. Just results.

Try Octokraft