Context Window Inspector

    1

    Full token breakdown of everything in your agent's context, sorted by size with keep/remove verdicts.

    Free

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Context Window Inspector

    Example session with this skill installed

    My responses feel degraded. /inspect-context — show me what's using up my context window.

    • Read your context and instructions
    • Compiled the context window inspector
    📊 Context snapshot — ~87,400 tokens used  |  ~112,600 remaining  |  44% full
    
    #  Source                          Type         Est. Tokens  % of Budget  Keep?
    1  Conversation (turns 1–18)       History      ~32,000      16.0%        ⚠️ Partial
    2  src/api/routes.ts               File         ~12,000       6.0%        ✅ Yes
    3  CLAUDE.md                       Instructions  ~4,200       2.1%        ✅ Yes
    4  node_modules/express/index.js   File          ~9,800       4.9%        ❌ No
    5  mcp: github (22 tools)          MCP           ~6,400       3.2%        ⚠️ Maybe
    
    🔥 Top bloat candidates
    - node_modules/express/index.js — loaded early, not referenced in current task
    - Turns 1–8 — initial setup discussion, no longer relevant
    - mcp: github — not used since turn 3
    
    ✅ Essential context
    - src/api/routes.ts, CLAUDE.md, turns 9–18
    
    💡 Quick wins
    - /clear 8 — removes first 8 turns (~14,000 tokens)
    - Close node_modules/express/index.js in your editor
    

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Produces a complete, sorted breakdown of everything currently loaded in your context window — system prompts, files, conversation history, skills, MCP tool definitions, and embedded documents — with token estimates and a keep/remove recommendation for each item.

    When to use

    • Before starting a complex task, to understand how much budget is already consumed
    • When responses feel degraded or shorter than expected
    • After a long session, to audit what has accumulated

    How it works

    Invoke it by name in your agent (in Claude Code: /inspect-context). It reads only what is already in the current session's context — no files are written, no network calls, no external services.

    Part of Claude Token Inspector, an open-source (MIT) suite of 8 context-window skills by the original author. Source: https://github.com/msishariful/claude-token-inspector

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for LLM Ops.

    Token counts are estimates (1 token ≈ 4 characters). Does not account for proprietary provider-side caching or hidden metadata.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free to use

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 23 days ago

    • Free with an Agensi account

    Permissions

    No special permissions declared or detected

    Listed23 days ago

    What's inside

    Frequently Asked Questions

    This skill identifies 'token bloat' by providing a diagnostic breakdown of what is filling your context window—such as bloated system prompts, large file attachments, or excessive tool definitions—helping you restore model focus and reduce API costs.

    It's a standard SKILL.md skill, built and tested on Claude Code. It also works in other SKILL.md-compatible agents (Cursor, Codex CLI, Gemini CLI, etc.), though suggested cleanup commands like /clear and /mcp are Claude Code conventions.

    No. Counts are estimates (about 1 token per 4 characters of English prose, slightly fewer for dense code). They're meant to show relative size and what to cut, not exact billing numbers.

    Yes, the skill is specifically designed to audit the overhead from connected MCP (Model Context Protocol) servers, identifying if specific tool definitions are consuming too much of your budget.

    The SKILL.md instructions file. It's free and open source (MIT); the same skill is part of the Claude Token Inspector repo on GitHub.

    New in LLM Ops

    Free