---
title: "01 sessionstart capy"
description: "SessionStart hook additional context: <context_window_protection> <priority_instructions> Raw tool output floods your context window. You MUST use capy MCP tools to keep raw data in the sandbox. </priority_instructions>"
type: skill
canonical_url: https://claudary.paisolsolutions.com/skills/01-sessionstart-capy
source: "Claudary"
difficulty: intermediate
author: "Claude Code Knowledge Pack"
date: 2026-07-10T10:58:42.779Z
license: CC-BY-4.0
attribution: "01 sessionstart capy — Claudary (https://claudary.paisolsolutions.com/skills/01-sessionstart-capy)"
---

# 01 sessionstart capy
SessionStart hook additional context: <context_window_protection> <priority_instructions> Raw tool output floods your context window. You MUST use capy MCP tools to keep raw data in the sandbox. </priority_instructions>

## Overview

SessionStart hook additional context: <context_window_protection>
  <priority_instructions>
    Raw tool output floods your context window. You MUST use capy
    MCP tools to keep raw data in the sandbox.
  </priority_instructions>

  <tool_selection_hierarchy>
    1. GATHER: capy_batch_execute(commands, queries)
       - Primary tool for research. Runs all commands, auto-indexes, and searches.
       - ONE call replaces many individual steps.
    2. FOLLOW-UP: capy_search(queries: ["q1", "q2", ...])
       - Use for all follow-up questions. ONE call, many queries.
    3. PROCESSING: capy_execute(language, code) | capy_execute_file(path, language, code)
       - Use for API calls, log analysis, and data processing.
  </tool_selection_hierarchy>

  <forbidden_actions>
    - DO NOT use Bash for commands producing >20 lines of output.
    - DO NOT use Read for analysis (use execute_file). Read IS correct for files you intend to Edit.
    - DO NOT use WebFetch (use capy_fetch_and_index instead).
    - Bash is ONLY for git/mkdir/rm/mv/navigation.
  </forbidden_actions>

  <output_constraints>
    <word_limit>Keep your final response under 500 words.</word_limit>
    <artifact_policy>
      Write artifacts (code, configs, PRDs) to FILES. NEVER return them as inline text.
      Return only: file path + 1-line description.
    </artifact_policy>
    <response_format>
      Your response must be a concise summary:
      - Actions taken (2-3 bullets)
      - File paths created/modified
      - Knowledge base source labels (so parent can search)
      - Key findings
    </response_format>
  </output_constraints>

  <capy_commands>
    When the user says "capy stats" or asks about context savings:
    → Call the capy_stats MCP tool and display the full output verbatim.

    When the user says "capy doctor" or asks to diagnose capy:
    → Call the capy_doctor MCP tool and display results as a checklist.
  </capy_commands>
</context_window_protection>

---

Source: [Claudary](https://claudary.paisolsolutions.com/skills/01-sessionstart-capy) · https://claudary.paisolsolutions.com