Know Your Code. Keep Your Context. Build Anywhere.
ContextSutra AI documentation
Everything needed to understand, use, audit, extend and publish ContextSutra, without asking anyone. Created by Vatsal Shah. Apache-2.0.
01. Executive summary
AI coding assistants are only as good as the project context they receive. Teams copy README files into prompts, switch between Claude Code, Cursor, Codex, Gemini CLI and
02. Functional specification
- Offline. No network access except the optional loopback server.
03. Architecture
text
04. Technical reference
bash
05. User guide
bash
06. Skills catalog
ContextSutra ships 12 Agent Skills in skills/portable/. Each is a SKILL.md with the same eight sections: Objective, When to use, Procedure, Never do, Contract, Worked exa
07. Security and legal
This page describes engineering controls. It is not legal advice and does not claim any compliance certification.
08. Testing and evidence
The suite uses Node's built-in test runner (npm test). It has two files: tests/core.test.mjs (capsule, redaction, freshness, install, server, CLI basics) and tests/v2
09. Audit and trust
npx contextsutra audit --out audit runs a fixed list of checks against the tool itself and writes audit.json, audit.md and audit.html. Every check shows its expected and
10. CI/CD and release
.github/workflows/ci.yml runs on every push and pull request across Ubuntu and Windows with Node 20, 22 and 24, and macOS with Node 24 (7 jobs; macOS runs Node 24 only).
11. FAQ and glossary
Does it need an API key or internet? No. It runs offline and never calls a model.
12. Contributor handbook
bash
13. Roadmap and status
1. Record real host activation runs for Claude Code and one other tool; update the compatibility matrix with versions and evidence.
14. Architecture decisions
Each record: context, decision, trade-offs, alternatives, status. All are accepted by the maintainer as of v0.2.0.
Proof
Raw evidence lives in evidence/: the full test run, sample reports in JSON, Markdown, HTML and JUnit, and the self-audit report.