SparseReading — AI skill for Claude Code
SparseRead: token-efficient reading for AI agents.
How to install SparseReading
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open Zedong-Liu/SparseReading and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What SparseReading does
SparseRead: token-efficient reading for AI agents. 读得更少，token 更省，证据不变。 Paper arXiv:2608.22237
Alternatives in AI
- Cursor Talk To Figma MCP — TalkToFigma: MCP integration between AI Agent (Cursor, Claude Code, Codex) and Figma, allowing Agentic AI to c 7k ★
- ARIS Agent Guide — For AI agents reading this repo 6.2k ★
- Clawcodex — Token efficient Claude Code full Python rebuild 875 ★
README
SparseRead
**Read less. Solve more.**
SparseRead is a training-free reading layer for tool-using agents. It controls which evidence enters the model context before an agent pays the cost of a broad read—while keeping provenance, refinement, verification, and native fallbacks explicit.
[](https://github.com/Zedong-Liu/SparseReading/actions/workflows/ci.yml) [](LICENSE) [Paper](https://arxiv.org/abs/2608.22237) · [Site](https://zedong-liu.github.io/SparseReading/) · [中文](README.zh-CN.md)
Agents are good at reasoning, but their default reading action is often still:
read everything -> put everything in context -> start reasoningThat is expensive for long reports, PDFs, workspaces, logs, spreadsheets, and multi-file audits. SparseRead adds a small control plane in front of native agent tools:
artifact -> Read Gate -> Reader Backend -> EvidencePack -> refine / verify / stopThe agent still decides what it needs. SparseRead makes the request bounded, source-anchored, and reversible when native access is the better path.
Results
The current paper evaluation covers 125 tasks, five workload scenarios, and six frontier models: Claude Opus 5, Qwen3.6-Plus, DeepSeek-V4-Flash, DeepSeek-V4-Pro, GLM-5.1, and Kimi-K2.5.
| Headline | Reported result |
|---|---|
| Maximum token reduction | 92.9% |
| Maximum wall-time reduction | 89.0% |
| Model–scenario cells with lower tokens and lower time | 30 / 30 |
| Cells preserving or improving task score | 26 / 30 |
| Sparse-fit cells preserving or improving task score | 22 / 24 |
The gain is not tied to one model: the evaluati
Related Skills
Quantum Reading Skill
Read Less, Understand More - Lightweight intelligent text reader for AI agents with 30%+ token optimization
Memory Lean
A token-lean MCP memory server for AI coding agents: a knowledge graph your agent can search without reading t
X402 Agent Search Examples
CyberWareX: web search, web page reading and token safety checks for AI agents, paid per call in USDC with x40
Claude Fable 5 System Prompt Clean
the optimized, token-efficient version of the leaked Claude Fable 5 / Mythos 5 system prompt. Re-engineered in
Pbi CLI
Power BI CLI - semantic models (.NET TOM) and PBIR reports for token-efficient AI agent usage, built for Claud
Browser39
A headless browser for AI agents that fetches modern web pages, runs JavaScript, manages sessions, and returns
Related Agents
Lean Coder
Token-efficient coding agent for any project. Use for edits, refactors and bug fixes (multi-file or unfamiliar
Metadata Extractor
Extracts paper metadata (authors, date, venue, fields, DOI/arxiv ID) and a paper-quality assessment (credibili
Deep Lit Reader
Read one arXiv paper in depth, write its wiki note, and emit a deep-lit result JSON.
