.cursorrules git:20260818.47cec68A

Immutable. This exact content is served forever at /api/v1/blob/785fdac4d7b60aaa.

---
name: agents
description: Always-loaded project anchor. Read this first. Contains project identity, non-negotiables, commands, and pointer to ROUTER.md for full context.
last_updated: [YYYY-MM-DD]
---

<!-- mex-tool-config: managed copy from .tool-configs/ -- keep this line so `mex check` can detect out-of-sync copies -->

# [Project Name]

## What This Is
<!-- One sentence. What does this project do?
     Length: 1 sentence maximum.
     Not a tagline — a factual description of what the software does.
     Example: "A REST API for managing inventory across multiple warehouse locations." -->

## Non-Negotiables
<!-- Hard rules the agent must never violate. Not preferences — rules.
     These are the things that, if broken, cause real damage to the codebase.
     Length: 3-7 items maximum. More than 7 means the list has not been prioritised.
     Example:
     - Never write database queries outside of the repository layer
     - Never commit secrets or API keys
     - Always handle errors explicitly — no silent failures -->

## Commands
<!-- The exact commands needed to work on this project.
     Include: run dev server, run tests, run linter, build.
     Use the actual commands from this codebase — not placeholders.
     Example:
     - Dev: `npm run dev`
     - Test: `npm test`
     - Lint: `npm run lint`
     - Build: `npm run build` -->

## Code Graph
The repo is indexed into `.mex/graph.db`. Use it to avoid re-reading code you already have — it is one tool alongside Grep/Glob, not a replacement for them.
- If you know the symbol name, go straight to it: `mex graph query <who-calls|what-calls|where-defined> <symbol>` and `mex graph get <id>` are exact and cheap. This is the strongest part of the graph. Give it exact names — an approximate name can return a confident wrong match.
- Exploring an unfamiliar task? `mex graph scope "<task>"` returns a compact JSONL manifest (`meta`, `fact`s, `summary`). Scope matches on words, not meaning: if your phrasing does not share vocabulary with the code, results will be weak. Treat it as a starting point, never as a complete answer.
- If the manifest does not clearly contain what you need, use Grep/Glob instead. Do not expand node ids that look irrelevant, and do not re-run `scope` with reworded phrasing more than once — that costs more than searching directly.
- Treat any source the graph DOES return as ALREADY READ; do not re-open those files.
- Pick 1-3 relevant node ids from the manifest and expand only those with `mex graph get <id> --detail source`.
- Before editing a symbol, run `mex impact <symbol|file>` to see affected callers and scaffold memory.
- If a result is `truncated`, do NOT repeat the broad query — narrow the task or use the summary's `suggestedNextCommands`. Scale through a few focused calls, never one giant response.
- During `mex sync`, adjudicate any AMBIGUOUS grounding; after repairs, ensure the refreshed grounding is re-emitted.

## After Every Task
After meaningful work, run GROW:
- Ground: what changed in reality?
- Record: update `.mex/ROUTER.md` and relevant `.mex/context/` files
- Orient: create or update a `.mex/patterns/` runbook if this can recur
- Write: bump `last_updated` on changed scaffold files and run `mex log` when rationale matters

## Navigation
At the start of every session, read `.mex/ROUTER.md` before doing anything else.
For full project context, patterns, and task guidance — everything is there.