cendor-init 0.2.1

Offline CLI to make your project Cendor-ready and Cendor-fluent for your AI assistant (init) and catch wiring mistakes (doctor). Run with `uvx cendor-init`.

cendor-init

One command to make your project Cendor-ready and Cendor-fluent for your AI assistant — plus a doctor that catches the wiring mistakes before they bite. Offline: no network, no API key.

uvx cendor-init            # detect + write assistant rules files (idempotent)
uvx cendor-init doctor     # validate the wiring; exit 1 on hard problems (CI-usable)

Optional developer tooling — dependency-free stdlib. It writes files and inspects your project; it makes no network call, and no Cendor library depends on it at runtime. (Node users: npx @cendor/init.) This is the Python twin of @cendor/init; both share the same behavior.

What init does

1. Detects your project — Python (pyproject.toml / requirements) or Node (package.json), which provider SDKs you have, and which cendor-* / @cendor/* packages are installed.

2. Writes the matching assistant rules file(s) so your assistant reads the correct Cendor call-shapes on every edit. Detected by default; --all for every one:

   • .github/copilot-instructions.md (GitHub Copilot)
   • .cursor/rules/cendor.mdc (Cursor)
   • AGENTS.md (the cross-tool default — always written)
   • a marked section in CLAUDE.md (Claude Code)
   • .windsurf/rules (Windsurf)

   Idempotent and safe: re-running updates a marker-delimited block in place — never duplicates, never clobbers your surrounding content. A dedicated file it didn't create is left alone unless you pass --force.

3. Offers MCP setup (--mcp) — drops the connect config for the Cendor MCP server (remote mcp.cendor.ai or local uvx cendor-mcp / npx @cendor/mcp) where it's absent.

4. Optional starter (--scaffold) — a minimal, correct starter for your language: a governed Agent loop (budget + guardrails + guard + audit) when cendor-sdk / @cendor/sdk is detected, otherwise an instrument() + budgeted-call example.

The rules content is a copy of section 3 of the docs source of truth, for-ai-assistants.

What doctor checks

Static checks only — it never mutates your project, and exits non-zero on hard problems so it works in CI:

• Namespace — a stray cendor/__init__.py in your tree, or a bare import cendor (the namespace has no module body — import from cendor.<tool>).
• Provider deps — a provider SDK your code imports but hasn't installed/declared (Cendor never pulls one for you; they are optional extras). Uses the installed environment when available.
• instrument() once — warns if Cendor is imported but the client is never wrapped.
• Money — flags coercing a price/cost to float (it should stay Decimal).
• Versions — warns when an installed/pinned cendor-* version trails the latest release.

uvx cendor-init --help

Options

Flag	Effect
--all	write every assistant rules file, not just the detected ones
--assistant <list>	comma-separated subset: copilot,cursor,agents,claude,windsurf
--mcp	also drop MCP connect config where absent
--scaffold	also write a correct starter — a governed Agent when the SDK is present, else instrument()+budget
--force	overwrite an owned file (.cursor/rules/cendor.mdc) even if not ours
--dry-run	show what would change without writing

Apache-2.0 · cendor.ai · For AI assistants · MCP