AI code assistant for legal-aware engineering — a Continue Fork tuned for production safety.
Claudio Legal is a fork of Continue, repackaged around a Claudio proxy and extended with workflows the original IDE extension does not ship: a pre-deploy audit, a workspace memory contract, curated MCP defaults, and tighter approval controls. Everything runs locally in VS Code; you point it at your own model gateway.
What's different from Continue
Vibe-to-Production Audit — one command scans your workspace for production blockers (missing auth, leaking secrets, unsafe SQL, untested critical paths) and lets the agent fix them in place.
Agent Memory Contract — a per-workspace contract of conventions, sensitive files, allowed commands, and forbidden actions. The agent reads it before acting, so rules survive across sessions. Auto-detects your stack (languages, frameworks, package manager, test runner) and refreshes on manifest changes.
Approval modes — explicit ask / plan / auto modes, plus a quick permissive toggle for trusted local work.
Default MCP servers — Context7 (live library docs) and Sequential Thinking are seeded on first run with no API key required. More templates ship as *.yaml.example; rename to enable.
~/.claudio config dir — coexists with stock Continue. Existing ~/.continue/config.yaml is migrated automatically when it points at the Claudio proxy.
Hardened Windows activation — CA setup is non-blocking and time-bounded, so corporate Windows machines never hang on crypt32.dll enumeration during extension load.
VS Code Agent
Agent works on development tasks alongside you, calling tools to read files, run commands, and edit code.
VS Code Chat
Chat to ask general questions and clarify code sections without leaving the IDE.
VS Code Edit
Edit to modify a code section in place without switching files.
VS Code Autocomplete
Autocomplete provides inline code suggestions as you type.
Quick start
Install the VSIX for your platform from extensions/vscode/build/.
Open the Claudio Legal sidebar; the tutorial file walks you through Autocomplete, Edit, Chat, and Agent.
Run Claudio: Vibe-to-Production Audit before your next deploy.
Run Claudio: View Agent Memory Contract to see what the agent inferred about your repo.
