Raker is a powerful CLI used by AI to review work against cloud-scale context. Built on Pinecone's Autocontext infrastructure, it acts as your comprehensive Reviewer Agent.
- Code Review: Apply organizational code intelligence to ensure commits align with internal best practices and your specific tech stack. Flag known anti-patterns, past bugs, and architectural deviations before they reach production.
- Design Review: Automatically review code and assets against your company's design system and brand guidelines to maintain visual and experiential consistency.
- Doc Review: Analyze extensive internal documentation to identify style deviations, factual inconsistencies, overlapping content, and potential plagiarism.
- Security Review: Proactively surface vulnerabilities and security risks by referencing historical incidents, unpatched flaws, and unstructured organizational security guidelines.
- Rust toolchain (cargo, rustc)
You can install the Raker CLI locally using the provided install script:
./scripts/install.shAlternatively, you can build and install it directly with Cargo:
cd cli
cargo install --path .Use the review command to review the contents of a directory or file:
# Review current directory
raker review
# Review a specific file
raker review src/main.rs(For more CLI commands, run raker --help)
We welcome contributions! Please review our Repository Guidelines (AGENTS.md) before getting started.
Before submitting a pull request, please ensure your changes pass all format and linting checks:
cd cli
cargo fmt --check
cargo clippy -- -D warnings
cargo testRAKER stands for Retrieval Augmented Knowledge Engine & Reviewer. It represents the core philosophy of this tool: an AI agent that retrieves organizational knowledge from Pinecone's cloud-scale context to intelligently evaluate and review code, designs, and documents before they enter the system.