skills.rest
agents-repo
agents-repoOFFICIAL
Explore all skills by this author →
WebsiteGitHub

maiconfz--context-token-reduction--context-token-chat

Analyze agent-context token waste from public project URLs, uploads, or pasted sources.

3|1|Updated May 2, 2026
Search Tags:#github#prompt optimization#ai agents#token analysis#context tokens#chat agent#footprint report
One-click install
npx skills add https://github.com/agents-repo/registry --skill maiconfz-context-token-reduction-context-token-chat-agents-repo
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: maiconfz--context-token-reduction--context-token-chat
Source: https://github.com/agents-repo/registry/tree/main/.cursor/skills/maiconfz/context-token-reduction/maiconfz--context-token-reduction--context-token-chat
Command: npx skills add https://github.com/agents-repo/registry --skill maiconfz-context-token-reduction-context-token-chat-agents-repo

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? AI agent configurations often waste context tokens through bloated always-on instructions, duplicated rules across targets, and missing ignore files. This Skill audits a project's agent-context footprint from a chat session—without needing local repository access—and reports where tokens are being wasted. ## Core Features & Use Cases - Remote Evidence Analysis: Fetches up to 15 key files (README, AGENTS.md, CLAUDE.md, Copilot/Cursor rules, ignore files) from public GitHub or git-forge URLs to assess context-token waste. - Structured Footprint Report: Emits a markdown report with findings grouped by dimensions such as always-on instructions, rules apply-mode, docs pulled into context, skills and packaged agents, ignore gaps, tree shape, and cross-target duplication, each with severity and token-effect labels. - Interview Fallback: When no usable URL, upload, or pasted source exists, it interviews the user with targeted questions instead of inventing repository facts. - Use Case: Paste a public GitHub repository URL into the chat and receive a severity-ranked report of context-token waste, plus a handoff pointing you to run the full reduction planning flow in an IDE. ## Quick Start Paste a public GitHub repository URL into the chat and ask it to analyze the project's agent-context token footprint.

Frequently Asked Questions about maiconfz--context-token-reduction--context-token-chat

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I analyze context token usage of a GitHub repository?▼

Paste a public GitHub repository URL into the chat. The agent fetches up to 15 key files such as README, AGENTS.md, CLAUDE.md, and rules files, then emits a footprint report grouping token-waste findings by dimension with severity ratings.

What files does a context token footprint analysis check?▼

It prioritizes instruction and context-load paths: README, CONTRIBUTING, AGENTS.md, CLAUDE.md, Copilot instructions, Cursor/Claude/Codex rules and skills, ignore files, and likely entry points. It prefers raw file URLs and does not clone or walk the whole tree.

Can it analyze private repositories or local files?▼

No. It only fetches public HTTPS URLs and never inspects a host working tree, clones repos, or uses credentials. For private projects, upload files or paste key file contents directly into the chat as evidence.

Why is the footprint report empty after I sent a URL?▼

The report stays empty when the URL is not usable evidence, such as a marketing page, a private repo, a 404, or when fetching is unavailable. In that case the agent interviews you or asks for uploads instead of inventing repository facts.

Does this agent create a token reduction plan?▼

No. It only reports findings and explicitly does not plan. After receiving a report, install the package in an IDE and run the reduce-context-tokens flow to generate an actual reduction plan.

Related Skills

Explore More Software Engineering Skills→
sameer-thakkarsameer-thakkar

create-agents-md

Generate AGENTS.md, CLAUDE.md, and .claude/rules/ files from codebase analysis.

Community
Advanced
cacheforge-aicacheforge-ai
10

context-engineer

Analyze workspace context files and generate token-usage optimization reports.

Official
Advanced
brianpunzalanbrianpunzalan

speckit-git-remote

Detect the GitHub owner and repository from a local Git remote URL.

Community
Basic
SelectXn00bSelectXn00b
198

model-usage

Track LLM token consumption and API costs from usage logs.

Official
Intermediate