Context Budget Analyzer

    1

    Analyzes where an Agent session's context budget goes: 7 buckets, token distribution, pollution and compaction risks, with deterministic parsing scripts.

    Free

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    context-budget-analyzer

    Example session with this skill installed

    Analyze my agent session's context usage / why is my context filling up

    • Read your context and instructions
    • Compiled the context-budget-analyzer

    A token budget breakdown across 7 categories with pollution risks and compaction recommendations

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

    About this skill

    Analyzes where an AI agent session's context budget actually goes. Categorizes session content into 7 buckets (system instructions, task goals, constraints, tool outputs, history, intermediate reasoning, state and memory), estimates token distribution, and identifies context pollution and compaction risks. Bundled Python scripts (estimate_tokens.py, analyze_session.py, analyze_usage.py) provide deterministic parsing of OpenClaw/JSONL transcripts. Works with Claude Code, OpenClaw, Codex and any agent that reads SKILL.md.

    Token counts are heuristic estimates. Does not modify session files. Requires Python 3 for script-based analysis. Specific model compression thresholds are not hardcoded.

    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 any payment.

    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 forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 14 days ago

    • Free with an Agensi account

    Permissions

    Terminal / Shell
    Read Files
    Listed14 days ago

    What's inside

    Frequently Asked Questions

    This skill detects 'token pollution' such as oversized tool outputs and redundant data that inflate costs and cause agents to 'forget' instructions. It provides a deterministic breakdown of how your context budget is being spent across seven distinct categories.

    The skill is designed for agents that export logs in JSONL format, Claude Code sessions, or OpenClaw trajectories. It uses standalone Python 3 scripts, making it compatible with any development environment that can output session transcripts or conversation logs.

    You will receive a suite of specialized Python 3 scripts for log parsing, a Markdown template for governance reporting, and a comprehensive guide on implementing the seven-category token budgeting framework.

    Unlike LLMs, which are notoriously poor at self-estimating token counts, this tool uses deterministic Python scripts to provide objective measurements. It identifies 'zombie reasoning' and constraint loss that manual prompting often overlooks.

    The scripts provide high-accuracy estimates based on character density and common encoding patterns. While they offer a precise view of distribution and decay, please note they do not directly modify your agent's internal code or session data.

    New in LLM Ops