dsh-deepread is a tool for DeepSeek Harness. Evidence-first deep reading for AI agents — trace claims, evidence, confidence and knowledge maps across articles, books and PDFs.
A well-documented, actively maintained evidence-first reading plugin with clear install paths and real example outputs w
Code quality8
README documents a structured dual-form architecture (portable Agent Skill plus host plugin) with a pure-JS PDF extractor, optional persistence/jobs/Web route, and batch comparison, but no test or CI details are visible.
Security7
No download-and-execute patterns, plaintext secrets, or excessive permissions are described; installs go through npm/npx and the dsh plugin CLI, though the actual permission surface is not shown.
Utility9
Addresses a real and common pain point by pairing claims with evidence, source locations, and four confidence levels, and explicitly refusing to fill gaps when sources lack support.
Maintenance8
Last updated 2026-08-22 with an unarchived repo, npm and GitHub release badges, and a published website, indicating ongoing activity.
Docs8
README provides two concrete install commands, a quick-start prompt example, a five-mode feature table, and links to reproducible real-output examples, though the install command field itself is listed as unknown.
Generated by AI after reading the project README, as a decision aid; neutral scores are given when information is thin. Not an official endorsement.
Evidence-first deep reading for AI agents — trace claims, evidence, confidence and knowledge maps across articles, books and PDFs.
How popular is dsh-deepread?
dsh-deepread has 39 stars and 3 forks on GitHub, last updated 2026-08-22.
What license does dsh-deepread use?
dsh-deepread is distributed under the MIT license.
Is dsh-deepread compatible with DeepSeek Harness?
dsh-deepread is listed in the DSH Universe directory as a tool for DeepSeek Harness. This plugin is listed in the DSH Universe directory and covered by its validation pipeline.
CLASSIFICATION EVIDENCE
/10类依据
项目类型技能
功能/10类文件与数据
规则置信度High
System reads GitHub Topics first, then compares against the in-site category dictionary and word-root rules. Current match:agent-skills、claude-code、document-analysis、pdf。
The first vision plugin for DeepSeek Harness, and the vision bridge for every text-only coding agent. Paste an image, get structured JSON evidence (OCR, layout, semantics). | The strongest vision add-on plugin for DeepSeek Harness, adding vision capability to text-only models like DeepSeek and GLM, paste an image and get structured JSON evidence (OCR, layout, semantics).
[dsh] A more powerful visual toolkit for text-only models: one-line install and use, paste images for direct recognition, multi-image Q&A, screenshot-to-frontend UI restoration, and more|DeepSeek Harness-native integration for agent-vision-toolkit: image Q&A, long-screenshot OCR, UI restoration, grounding, pixel diff, Artifacts, and Web UI.