MCP servers and skills: search, filter by where they run, and compare.
Experimental harness for measuring information fidelity across MCP, A2A, and A2UI layers — companion repo for the Fidelity of the Agentic Stack paper
SkillMITFew starsNo key
An unmaintained reference skill for fidelity-locked implementation of frozen UI designs using canonical visual shells, dynamic islands, and visual regression.
SkillApache-2.0Few starsNo key
A Claude skill for translating documents where mistranslation has consequences — terminology consistency, self-critique on safety-critical content, back-translation verification.
SkillMIT★ 2Few starsNo key
Agent Skill for recovering incomplete Figma evidence before high-fidelity implementation
SkillMIT★ 5No key
Preserve fidelity across prototype iterations: freeze the baseline, bound changes, reuse real assets, verify in-browser.
SkillMITFew starsNo key
高还原 UI 协作 skill：Figma / 截图还原、视觉约束、动效排查与最小改动。Portable agent skill for faithful UI implementation and evidence-based debugging.
SkillMIT★ 3Few starsNo key
蒸馏投毒评测skill，用于评测一个修改后或投毒后的 skill，是否仍然保留 clean skill 的能力、推理方式、边界完整性与风格辨识度。
SkillMIT★ 1Few starsNo key
Task Fidelity｜任务执行防跑偏：贯通需求理解、设计、执行核对与结果验证的 Agent Skill
SkillMITFew starsNo key
Check that a rewrite kept the meaning: numbers, literals, conditions, thresholds (at least vs at most), comparison direction, claim strength, and modal strength (MUST/SHOULD/MAY, 必须/建议/可以). English +
SkillMIT★ 4Few starsNo key
Claude skills for generating ASCII components and screens for humans to review, and AI to build from.
SkillMIT★ 17No key
The handshake for agents acting on the open internet. Six layers say the authorization is legal; the seventh asks the only question that matters once it is — is the agent still you?
SkillMIT★ 1Few starsNo key
根据产品需求生成统一规范的 Web 端与移动端单 HTML 高保真原型，并提供 UI 组件规范建设指引
SkillMIT★ 7No key
Audits frontend implementations for design-system drift across CSS, Tailwind, JSX, TSX, Vue, and Angular code.
SkillMITFew starsNo key
Self-contained Codex skill for high-fidelity PDF to editable Word conversion: reference formatting, native equations, three-line tables, TOC and verification.
SkillMIT★ 3Few starsNo key
Check that an adaptation of a source text stays faithful: verbatim quotes, no invented sentences or numbers, bounded framing, declared omissions.
SkillMIT★ 1Few starsNo key
Memory washer/defragmenter for AI agent memory: class-B semantic wash (fidelity-gated, blocks any structured-token drop, snapshot+rollback) and class-A byte-identity session estate (ULTRA archive/rest
SkillApache-2.0★ 11No key
A modular high-fidelity writing editor combining voice preservation, factual fidelity, adaptive rewriting, genre-aware editing, pattern analysis and long-document refinement.
SkillMIT★ 1Few starsNo key
Vendor-neutral Agent Orchestrator for developer workflows with LangGraph, AGENTS.md, and multi-agent pipelines
SkillMIT★ 3Few starsNo key
Claude Code plugins that make Claude build like a senior specialist. high-fidelity-web: awwwards-grade animated marketing/landing sites (design tokens, motion, procedural web 3D). rails-testing: senio
SkillMIT★ 1Few starsNo key
蒸馏工厂：把书籍/长视频/播客/访谈/人物素材蒸馏成可执行、可验证、可追溯的 Agent Skills。contract-first + FIDELITY 双 Agent 出厂质检 + 跨模型异族复评（DeepSeek，组织独立台账可查）。五道机器门禁不过不发版；82 个自身缺陷公开登记，数字由测试强制相等。零第三方依赖，默认离线。
SkillMIT★ 2Few starsNo key
This repository includes code and materials for the paper "SiliconBench: Speed, Memory, and Fidelity for LLM Serving on Unified-Memory Desktops"
SkillMIT★ 8No key
End-to-end Micro-Expression Recognition paper reproduction skill for Claude Code & Codex CLI. PDF -> faithful PyTorch -> 13 standard experiments across 7 MER datasets.
SkillMIT★ 8No key
High-fidelity Office document skills for Codex across Word, Excel, and PowerPoint.
SkillMIT★ 3Few starsNo key
Claude Code plugin: design-first TDD for UI components
SkillMITFew starsNo key
Who stands behind it
