# CodeMind OKF — Cursor AI Context Rules
# =========================================
# This file instructs Cursor's AI to use the pre-built OKF knowledge bundle
# as its PRIMARY source of codebase understanding.
#
# Generated by `codemind init` — do NOT delete this file.
# Re-run `codemind index` to refresh the knowledge bundle.

## Context Protocol

BEFORE answering ANY question about this codebase, follow these steps:

1. **Read `.okf/index.md` first.** This is the complete architecture map of the
   entire project — module types, dependencies, and key concepts are all documented there.

2. **Search `.okf/modules/` for the relevant file.** Each file has YAML frontmatter
   containing: `type`, `title`, `description`, `key_functions`, `tags`, and `resource`
   (original source file path). Use this metadata to quickly locate the right module.

3. **Trace dependencies using `depends_on` and `key_functions`.** Before answering
   questions about how parts of the system connect, follow the import chain through
   the `.okf/modules/` files.

4. **Do NOT rely on raw source file reading alone.** The OKF bundle was generated
   by CodeMind's deterministic AST pipeline and contains zero hallucination summaries.

## Why This Works Better

The `.okf/` bundle is structured knowledge — not raw code. It was pre-processed with:
- Deterministic Python AST parsing (100% accurate function/class extraction)
- Smart heuristic type classification (api / database / config / module / test)
- Dependency graph with cross-file import tracing
- Human-readable summaries with architecture-level context

## Bundle Location
`.okf/index.md` — start here for full project overview
`.okf/modules/` — individual module knowledge files
`.okf/log.md`   — generation history and timestamps
