# CodeVetter docs

> Local-first knowledge system for CodeVetter — the AI desktop code review workbench for agent-generated code.

## Docs

- [CodeVetter docs](https://codevetter.com/docs/): The local-first knowledge system for the CodeVetter repository.
- [Business-rule archaeology](https://codevetter.com/docs/BUSINESS_RULE_ARCHAEOLOGY/)
- [Business-rule archaeology gates](https://codevetter.com/docs/BUSINESS_RULE_ARCHAEOLOGY_QUALIFICATION/)
- [Local differential verification](https://codevetter.com/docs/DIFFERENTIAL-VERIFICATION/)
- [Local History Explorer](https://codevetter.com/docs/HISTORY-EXPLORER/)
- [Scenario compilation](https://codevetter.com/docs/SCENARIO-COMPILATION/)
- [Structural graph coverage contract](https://codevetter.com/docs/STRUCTURAL-GRAPH-COVERAGE/)
- [Warm local verification](https://codevetter.com/docs/WARM-VERIFICATION/)

## Architecture

- [How CodeVetter works: end-to-end](https://codevetter.com/docs/architecture/how-it-works/): Learning-tier entry point for the native app, shared Rust authority, CLI, MCP, and evidence-backed verification loop.
- [Architecture overview](https://codevetter.com/docs/architecture/overview/): How the native macOS app and Rust verification core are layered.
- [Native boundary and command map](https://codevetter.com/docs/architecture/ipc-and-commands/): How the SwiftUI app, CLI, and MCP server share the Rust verification core.
- [Data model](https://codevetter.com/docs/architecture/data-model/): SQLite tables, persistence boundaries, and what lives where.
- [Review pipeline](https://codevetter.com/docs/architecture/review-pipeline/): The review → fix → re-review → proof flow and how findings are produced.
- [Structural graph and history workbench](https://codevetter.com/docs/architecture/graph-and-history/): The canonical syntax-aware graph, release-history reconstruction, and their analytics boundary.
- [T-Rex MCP projection](https://codevetter.com/docs/architecture/trex-mcp-projection/): The authorization and receipt contract for a future agent-triggered change-preview verifier.
- [Local evidence MCP](https://codevetter.com/docs/architecture/mcp-sidecar/): Opt-in, read-only, stdio-only MCP server exposing one repository's qualified graph, history, review, and business-rule evidence.
- [Local history evidence imports](https://codevetter.com/docs/architecture/history-evidence-import/): Attach provider-side outcomes to local repository history without calling a hosted API.
- [Native macOS migration matrix](https://codevetter.com/docs/architecture/native-migration-matrix/): The no-regression contract for transferring every current CodeVetter surface to the native Evidence Workbench.
- [Native Rust boundary](https://codevetter.com/docs/architecture/native-rust-boundary/): Evidence-backed ownership split between the macOS process and supervised Rust workers.
- [Repo Unpack](https://codevetter.com/docs/architecture/repo-unpacked/): Evidence-backed repository snapshots, briefs, graph, history, and exports.
- [Verification workbench evidence](https://codevetter.com/docs/architecture/verification-workbench/): Additive identity, lifecycle, retention, and performance records that connect CodeVetter's existing specialist surfaces.

### Decisions

- [Decision: MCP Rust SDK](https://codevetter.com/docs/architecture/decisions/mcp-sdk/): Chose the official modelcontextprotocol/rust-sdk crate rmcp pinned to 2.2.0 for the local history sidecar.
- [Decision: OSS integration evaluation](https://codevetter.com/docs/architecture/decisions/oss-integration/): Evaluation of OSS integrations that strengthen CodeVetter's local evidence loop.
- [Decision: Structural graph contract](https://codevetter.com/docs/architecture/decisions/structural-graph-contract/): The release-qualified capability, trust, interchange, and performance contract for CodeVetter's local code graph.

## Development

- [Local setup](https://codevetter.com/docs/development/setup/): Prerequisites, install, and how to build CodeVetter's native app, Rust core, and landing page.
- [Testing](https://codevetter.com/docs/development/testing/): Test, qualification, and code-health surfaces for the native app and shared Rust core.
- [Configuration](https://codevetter.com/docs/development/configuration/): Native build identity, local settings, data ownership, and release-only protected inputs.
- [Working on docs](https://codevetter.com/docs/development/docs/): How the docs tree is organized, validated, and rendered with Blume.
- [Verification receipts](https://codevetter.com/docs/development/verification-receipts/): Machine-first ingestion and comparison of project-owned test and performance evidence.
- [Agent-task corpus contracts](https://codevetter.com/docs/development/agent-task-corpus/): Immutable task-package contracts and fail-closed readiness checks for the coding-agent corpus.
- [Catch-rate benchmark](https://codevetter.com/docs/development/benchmark/): Starter harness for measuring whether CodeVetter catches known bugs in agent-generated changes.
- [Native macOS development](https://codevetter.com/docs/development/native-macos/): Reproducible tooling and ownership boundaries for the AppKit and SwiftUI Evidence Workbench.
- [Performance harness and baselines](https://codevetter.com/docs/development/performance/): How CodeVetter measures native responsiveness, Rust workloads, and regression claims.

## Knowledge

- [Failed approaches and reusable mistakes](https://codevetter.com/docs/knowledge/failed-approaches/): Things that broke, why, and what not to try again.
- [Competitive landscape](https://codevetter.com/docs/knowledge/competitive-landscape/): How CodeVetter is positioned against hosted PR-review bots, agent orchestrators, and Claude Code /review.
- [Third-party tooling decisions](https://codevetter.com/docs/knowledge/tooling-decisions/): What CodeVetter embeds instead of building, what it refuses, and why — the hub page for per-category evaluations.
- [Secret scanning and supply-chain tooling](https://codevetter.com/docs/knowledge/tooling-secrets-and-supply-chain/): Evaluation of secret scanners and dependency/vulnerability scanners against CodeVetter's offline and privacy constraints.
- [Code coverage tooling](https://codevetter.com/docs/knowledge/tooling-coverage/): Coverage collection across Rust, TypeScript and Python, the LCOV transport decision, and the two traps that produce false verdicts.
- [Agent benchmark corpora](https://codevetter.com/docs/knowledge/tooling-agent-benchmarks/): Evaluation of public agent/web benchmarks for licensing, deterministic grading, and offline reproducibility.
- [Documentation tooling — the DeepWiki boundary](https://codevetter.com/docs/knowledge/tooling-documentation/): Where DeepWiki helps CodeVetter maintainers, where it is rejected, and what would fit for generated architecture docs.
- [Sandboxed execution and static analysis](https://codevetter.com/docs/knowledge/tooling-sandboxing/): Isolation options for running untrusted agent code on Apple Silicon, determinism controls, SARIF-emitting analyzers, and the CodeQL licensing blocker.
- [Codebase context tools landscape](https://codevetter.com/docs/knowledge/codebase-context-tools-landscape/): Survey of codebase indexing, context-provider, and documentation tools evaluated for CodeVetter (research snapshot, April 2026).
- [Runtime performance lab publication record](https://codevetter.com/docs/knowledge/performance-lab-publication/): Published product experiments, external contributions, evidence strength, and local-cost boundaries from the first CodeVetter performance campaign.

### Learnings

- [Learning roadmap](https://codevetter.com/docs/knowledge/learnings/README/): Every non-obvious concept this project runs on, with one canonical home each.
- [Learnings: platform and stack](https://codevetter.com/docs/knowledge/learnings/new-things/): How CodeVetter uses Tauri, typed IPC, CLI-agent subprocesses, structural scanning, local storage, provider APIs, updates, telemetry, and release tooling.
- [Learnings: telemetry and indexing](https://codevetter.com/docs/knowledge/learnings/telemetry-and-indexing/): How CodeVetter parses agent transcripts, deduplicates usage, tracks incremental cursors, calculates costs and quotas, and indexes history in the background.
- [Learnings: verification and judgment](https://codevetter.com/docs/knowledge/learnings/verification-and-judgment/): How CodeVetter judges candidate changes with staged verification, provenance, deterministic evidence grades, isolated worktrees, and synthetic QA.

## Operations

- [Automation readiness contract](https://codevetter.com/docs/operations/automation-contract/): The privacy-safe product, release, reliability, and Foundry evidence contracts for CodeVetter.
- [Release pipeline](https://codevetter.com/docs/operations/release-pipeline/): How a native version bump becomes a signed and notarized GitHub Release.
- [Landing page deploy](https://codevetter.com/docs/operations/landing-deploy/): How the Astro landing page deploys to Cloudflare Pages.
- [CI](https://codevetter.com/docs/operations/ci/): GitHub Actions checks for the native app, Rust core, CLI, MCP, site, and protected release.
- [Signing keys and auto-update](https://codevetter.com/docs/operations/signing-and-updates/): Developer ID signing, notarization, and Sparkle appcast inputs for the native macOS app.

### Jobs

- [Weekly quality check](https://codevetter.com/docs/operations/jobs/weekly-quality/): The Monday cron that runs available quality scripts across the repo.

### Runbooks

- [Cut a desktop release](https://codevetter.com/docs/operations/runbooks/cut-a-release/): Runbook for shipping the signed and notarized native CodeVetter app.
- [Deploy the landing page](https://codevetter.com/docs/operations/runbooks/deploy-landing/): Runbook for deploying the Astro site to Cloudflare Pages.

## Product

- [Product overview](https://codevetter.com/docs/product/overview/): What CodeVetter is, the durable scope, and the current capability matrix.
- [Surfaces and navigation](https://codevetter.com/docs/product/surfaces/): The six native macOS sections and their shared CLI and MCP contracts.
- [T-Rex change and preview verification](https://codevetter.com/docs/product/trex-change-preview/): Run bounded browser verification for an exact pull request or commit range against an existing preview.
- [Capability glossary](https://codevetter.com/docs/product/capabilities/): What CodeVetter uses today, where each capability is available, and what is intentionally future work.
- [Synthetic user QA](https://codevetter.com/docs/product/synthetic-user-qa/): The runtime proof layer for agent-written code: exercise a changed web surface and store evidence next to review findings.