README.md

Documentation

The root README is the short product introduction. Use this map when you need the operating contract or policy details behind it.

From a managed repository, the installed CLI reads the documentation carried by that repository's locked release:

code-polishy docs list
code-polishy docs find behavior review
code-polishy docs read agent-workflows
code-polishy docs read adding-a-language

The checked-in catalog.json owns stable topic identifiers, titles, summaries, and aliases. Search is local, bounded, and deterministic.

Start here

• Installation explains archive bootstrap, exact-source recovery, two-phase upgrades, OCI execution, manifests, and target locks.
• AI-Agent Setup and Adoption is the authoritative procedural runbook for setup and upgrades performed by a coding agent.
• Adopting Code Polishy is the detailed project-modeling reference for modules, capabilities, checks, tests, and CI.

Operating Code Polishy

• Agent Workflows is the short mandatory entry point for ordinary work, event-selected verification, and conditional procedures.
• Agent Reviews defines boundaries for selected reviews.
• Machine-readable Reports defines the canonical verdict and complete results for automation.
• Isolated Task Sessions specifies disposable-worktree boundaries, promotion, and artifacts.
• Artifact Security documents shared container scanning, producer contracts, OpenVEX, and optional behavior-review evidence custody.
• Repository Size explains workspace and governed-content footprint reports, Git-base comparisons, and agent interpretation.
• Repository Trends explains weekly rewrite, copy, change-spread, revert, and flaky-test trends.

Policy reference

• Architecture
• Behavior and final-state review, including experimental opt-in feature policy and task requests
• Code quality
• Conditional modules
• End-to-end reliability
• Exceptions
• Portability and external inputs
• Security
• Supply chain
• Test strength
• Verification and testing

Maintainer reference

• Dependency Assessment Decisions explains applicability, agent ownership, and the risk comparison before early admission.
• Security Evidence Lifetime explains bounded scan reuse and security refreshes without repeating source verification.
• Current JavaScript Dependency Assessments records exact advisory call paths and scheduled remediation for this repository.
• Canonical Agent Guidance Design defines what belongs in managed AGENTS.md, where longer rationale lives, and how to update the template.
• Repository Bootstrap Wrapper defines the one-command onboarding path, source trust boundary, and managed adoption ownership.
• Behavior Review Intent Custody explains conditional request capture, stdin transport, and explicit evidence cleanup.
• Repository Size Analysis defines safe, bounded footprint measurement and its human/agent interpretation boundary.
• Repository Trends Design explains history replay, landing semantics, gate-record reading, and why trends exclude enforced measures.
• Reliability Reminder defines scoped advisory selection, rendering, and its non-enforcement boundary.
• Policy Engine Architecture maps the runtime ownership boundaries and execution model.
• Source Comment Policy Design explains the default-preserving switch and the rationale for repositories that select the strict no-comment boundary.
• Rust and Shader Input Coverage defines Rust control inputs, WGSL ownership, and the current analysis boundary.
• Native Rust Test Evidence defines Cargo selectors, managed Nextest results, retries and reusable test receipts.
• Rust Support Plan tracks the beta implementation and the remaining requirements for first-class Rust support.
• Source Provenance records public design references, toolchain origins, and implementation boundaries.
• Release Checklist defines immutable tags and the combined native-archive and GHCR publication workflow.
• Adding a Language Pack defines local language packs, conformance verification, exact installation, and adapter protocol.