---
description: context-mode routing rules for context window protection
alwaysApply: true
---

# context-mode

Raw tool output floods context window. Use context-mode MCP tools to keep raw data in sandbox.

## Think in Code — MANDATORY

Analyze/count/filter/compare/search/parse/transform data: **write code** via `ctx_execute(language, code)`, `console.log()` only the answer. Do NOT read raw data into context. PROGRAM the analysis, not COMPUTE it. Pure JavaScript — Node.js built-ins only (`fs`, `path`, `child_process`). `try/catch`, handle `null`/`undefined`. One script replaces ten tool calls.

## Tool Selection

0. **MEMORY**: `ctx_search(sort: "timeline")` — after resume, check prior context before asking user.
1. **GATHER**: `ctx_batch_execute(commands, queries)` — runs all commands, auto-indexes, searches. ONE call replaces many steps.
2. **FOLLOW-UP**: `ctx_search(queries: ["q1", "q2", ...])` — all follow-up questions, ONE call (default relevance mode).
3. **PROCESSING**: `ctx_execute(language, code)` | `ctx_execute_file(path, language, code)` — sandbox, only stdout enters context.
4. **WEB**: `ctx_fetch_and_index(url)` then `ctx_search(queries)` — never dump raw HTML.
5. **INDEX**: `ctx_index(content, source)` — store in FTS5 for later search.

## Parallel I/O batches

For multi-URL fetches or multi-API calls, **always** include `concurrency: N` (1-8):

- `ctx_batch_execute(commands: [3+ network commands], concurrency: 5)` — gh, curl, dig, docker inspect, multi-region cloud queries
- `ctx_fetch_and_index(requests: [{url, source}, ...], concurrency: 5)` — multi-URL batch fetch

**Use concurrency 4-8** for I/O-bound work (network calls, API queries). **Keep concurrency 1** for CPU-bound (npm test, build, lint) or commands sharing state (ports, lock files, same-repo writes).

GitHub API rate-limit: cap at 4 for `gh` calls.

## Forbidden Actions

- DO NOT use Bash for >20 lines output — use `ctx_execute` or `ctx_batch_execute`.
- DO NOT use Read for analysis — use `ctx_execute_file`. Read IS correct for Edit.
- DO NOT use WebFetch — use `ctx_fetch_and_index`.
- DO NOT use curl/wget in terminal — use `ctx_fetch_and_index`.
- Bash ONLY for git, mkdir, rm, mv, navigation, short commands.
- DO NOT use `ctx_execute`/`ctx_execute_file` to create/modify files. ctx_execute is for analysis and computation only.

## File Writing Policy

ALWAYS use native file editing tools to create/modify files. NEVER use `ctx_execute`, `ctx_execute_file`, or Bash to write file content.

## Output

Write artifacts to FILES — never inline. Return: file path + 1-line description.

## Session Continuity

Skills, roles, and decisions persist for the entire session. Do not abandon them as the conversation grows.

## Memory

Session history is persistent and searchable. On resume, search BEFORE asking the user:

| Need | Command |
|------|---------|
| What did we decide? | `ctx_search(queries: ["decision"], source: "decision", sort: "timeline")` |
| What constraints exist? | `ctx_search(queries: ["constraint"], source: "constraint")` |

DO NOT ask "what were we working on?" — SEARCH FIRST.
If search returns 0 results, proceed as a fresh session.

## ctx Commands

| Command | Action |
|---------|--------|
| `ctx stats` | Call ctx_stats MCP tool, display full output verbatim. |
| `ctx doctor` | Call ctx_doctor MCP tool, run returned shell command, display as checklist. |
| `ctx upgrade` | Call ctx_upgrade MCP tool, run returned shell command, display as checklist. |
| `ctx purge` | Call ctx_purge MCP tool with confirm: true. Warn user this is irreversible. |

After /clear or /compact: knowledge base and session stats preserved. Use `ctx purge` to start fresh.