Caveman is a multi-agent AI coding skill/plugin that compresses LLM output tokens by ~65-75% using a 'caveman' terse communication style. It's not a library or service but a distributed configuration package that installs into 30+ AI coding agents (Claude Code, Codex, Gemini, Cursor, Windsurf, Cline, Copilot, opencode, OpenClaw, etc.) through their native plugin/extension/rule-file mechanisms. The core deliverable is a set of SKILL.md prompt files that instruct models to drop articles, filler, pleasantries, and hedging while preserving exact technical substance, code blocks, URLs, and paths. The architecture is a single Node.js installer (bin/install.js) that auto-detects installed agents and dispatches to per-provider installation paths. For Claude Code specifically, it adds SessionStart and UserPromptSubmit hooks plus a statusline badge. The repo includes several sub-skills: caveman-commit (terse conventional commits), caveman-review (one-line PR comments), caveman-compress (rewrites memory files like CLAUDE.md into caveman format to save input tokens), caveman-stats (reads session logs to report real token savings), and cavecrew (compressed subagent definitions for investigator/builder/reviewer patterns). There's also caveman-shrink, an MCP middleware proxy that compresses tool descriptions. The codebase demonstrates strong engineering practices: symlink-safe flag file writes with O_NOFOLLOW and ownership verification, atomic settings.json merges with JSONC tolerance, defensive hook validation before writes, idempotent install/uninstall cycles, and a three-arm eval harness (baseline/terse/skill) for honest token compression measurement. The Python compress tool has safety guards against exfiltrating sensitive files and verifies backup integrity before overwriting. Test coverage is substantial across installer e2e flows, hook symlink safety, opencode plugin lifecycle, and config resolution order.
This repository is a knowledge package designed to create an AI agent skill that embodies Charlie Munger's thinking framework. It doesn't contain executable code but rather curates and structures Munger's mental models, decision heuristics, and communication style into a usable format for AI agent runtimes. The core output is the SKILL.md file, which provides detailed instructions for how an AI agent should adopt Munger's perspective when responding to user queries, including role-playing rules, a response workflow requiring research before answering, and specific guidelines for expressing Munger's characteristic communication style. The repository represents a sophisticated approach to expertise capture and encoding for AI systems. It synthesizes information from over 50 sources including Berkshire Hathaway shareholder meetings, USC/Harvard speeches, interview records, and external criticisms to distill Munger's approach into five core mental models, eight decision heuristics, and a detailed expression DNA. The references/ directory contains the source research materials, demonstrating transparency about the knowledge foundation. Examples/show how the skill works in practice through sample conversations about investment decisions, team management, and content creation. While primarily a knowledge resource, the structure of the SKILL.md file itself offers potential reusability as a template for creating similar 'thinking style' skills for other experts or domains. The systematic approach to encoding expertise through role-playing protocols, mental model collections, decision frameworks, and communication guidelines represents a pattern that could be adapted with domain-specific content substitution for other AI agent skills.
This repository is a distribution and development monorepo for "Agent Skills" — packaged instruction sets, rule files, and CLI tooling that extend AI coding agents to work more effectively with Vercel, React, Next.js, and React Native. The dominant artifact is the `vercel-optimize` skill, a sophisticated performance/cost audit pipeline. It starts by collecting real signals from the Vercel CLI (metrics, usage, project config, route data), then applies deterministic pure-JS gates and AST-based code scanners to generate candidate investigation targets. LLM sub-agents receive structured investigation briefs with playbooks, citations, and support-topic guardrails; their outputs are collected, mechanically verified, sanitized, deduplicated, graded, and rendered into a customer-safe markdown report. This is not a web service or library in the traditional sense — it is a toolchain of executable scripts plus knowledge artifacts designed to be installed into agent environments via `npx skills add`. The repository also contains curated knowledge bases (`react-best-practices`, `react-native-skills`, `composition-patterns`, `react-view-transitions`, `web-design-guidelines`) stored as markdown rules with YAML frontmatter. A build package (`react-best-practices-build`) parses these rules, validates them, and compiles them into `AGENTS.md` files for consumption by agents. The engineering quality is strongest in `vercel-optimize`: it has extensive `node:test` coverage, real CLI output fixtures, a modular gate/scanner/sanitizer architecture, claim-based verification, and careful handling of Vercel API rate limits. The main risks are external dependency drift — Vercel CLI output schemas, metric IDs, Observability Plus availability, and Next.js/React framework versions all change over time — and the inherent fragility of parsing LLM sub-agent output. Although the repo contains large amounts of markdown knowledge content, the dominant character is application/tooling: it ships executable entry points, has unit tests, CI workflows, and a build pipeline. It is not a data-research repository, and the metric fixtures exist to test the tooling rather than to answer analytical questions.
OPC Skills is a curated marketplace and toolchain of AI agent skills designed for solopreneurs, indie hackers, and one-person companies. The repository contains 9 skills (reddit, twitter, producthunt, domain-hunter, logo-creator, banner-creator, nanobanana, seo-geo, requesthunt, archive) that each provide structured prompts and Python scripts to automate tasks like market research, SEO/GEO optimization, domain hunting, logo/banner creation, and social media analysis. Skills are distributed via the Agent Skills standard (`npx skills add`) and also available as a Claude Code plugin. The project includes a full website at opc.dev powered by a Cloudflare Worker (website/worker.js) with per-skill pages, a blog, SEO/GEO optimization, and install statistics tracking. CI/CD pipelines deploy the website, test installs across platforms, and fetch install stats daily. The codebase follows a consistent architecture: each skill has a SKILL.md with YAML frontmatter, optional Python scripts following a credential → API wrapper → per-endpoint script pattern with TOON (Token-Oriented Object Notation) output, and dependencies declared between skills. While primarily an application, it has data-research characteristics due to its heavy use of external API scraping for market signals (Reddit, Twitter, ProductHunt) and SEO analytics. The repository is actively maintained with frequent releases, a changelog, and an agent-assisted development workflow (Ralph). Notable strengths include the reusable skill template pattern, cross-session memory via the archive skill, and the standalone SEO audit script. Key risks include reliance on multiple third-party API keys, tight coupling between some skills, and fragility in the install stats scraper.
This repository provides a generic-based dependency injection framework for Go, called 'kod' (Killer Of Dependency), along with a CLI tool for code generation, component registration, and dependency graph validation. It features a rich interceptor chain for cross-cutting concerns (retry, validation, tracing, rate limiting, circuit breakers, etc.), and includes a comprehensive test suite and examples. The codebase also incorporates an extensive set of AI agent skills for software development workflows (grilling, code review, TDD, domain modeling) from the mattpocock/skills ecosystem. The primary value is the DI framework that leverages Go generics to provide compile-time safety for component wiring, with generated boilerplate code. The architecture is modular with clear separation of core, interceptors, internal utilities, and tests. It is actively maintained with modern Go features.
This repository is a comprehensive, tool-agnostic skills and steering pack that teaches AI coding agents to perform AWS Well-Architected Framework reviews. It contains a curated reference corpus of over 2,000 markdown files (6 pillars, 307 best practices, 27 industry lenses) derived from official AWS documentation, plus per-tool adapters for 15 AI coding assistants (Claude Code, Cursor, Codex, GitHub Copilot, etc.). The repo also ships production-grade supporting tooling: a CI gating tool (wa-ci) that diffs structured review JSON against a baseline and blocks PRs on regressed gaps; an evaluation harness with blind multi-reviewer adversarial adjudication; a crawler to regenerate the reference corpus from AWS docs; and one-command installers. The core skill dispatches six parallel subagents (one per pillar) to keep the 2.2 MB corpus out of any single context window. Despite the extensive content, the repository has notable maintenance challenges: the reference snapshot will drift as AWS documentation evolves (users are responsible for regeneration); lens files are duplicated across `powers/` and `skills/` directories creating a sync hazard; and the flagship skill is explicitly deprecated in favor of a successor in Agent Toolkit for AWS. The evaluation harnesses and CI tooling are well-crafted and reusable, but the sheer scale of markdown files and platform-specific adapters makes ongoing synchronization complex.
Agentic Awesome Skills is a curated marketplace infrastructure for AI coding agent skills—instructional playbooks (SKILL.md files) with accompanying scripts and assets. The repository provides a multi-platform distribution system supporting Claude Code, Cursor, Gemini CLI, Codex CLI, Kiro, and others. It includes a React+Vite web application for browsing, searching, and installing skills; a CLI and MCP server for local agent integration; and a comprehensive validation, security scanning, and release pipeline. The codebase implements a transaction engine with deterministic content addressing (canonical JSON + SHA-256), durable state management, and schema-validated skill installation, forming a robust foundation for agent-owned tool selection and workflow planning. While the core infrastructure (AAS v1 library, MCP server, catalog search) demonstrates high craftsmanship and reusability, the repository is hampered by massive file duplication: identical skill scripts and governance modules appear under `skills/`, `plugins/agentic-awesome-skills/`, and `plugins/agentic-awesome-skills-claude/`, creating synchronization risk. Some skills contain hardcoded paths or incomplete boilerplate. The repository is actively maintained with CI/CD, automated audits, and security scrutiny, though the sheer volume of 2,474+ skills imposes ongoing maintenance burden. Overall, it is a well-engineered, production-quality skill distribution platform with notable reusable design patterns.
This repository is a motion-graphics video production toolkit designed for AI coding assistants (Codex and Claude Code). It provides a structured pipeline for turning talking-head footage into edited videos with animated overlays, using HTML/CSS/JS compositions driven by GSAP timelines. The system includes reusable style libraries (Vox Explainer and Kallaway) with hundreds of parameterized card templates, transcription and cut-detection scripts, beat-planning heuristics, and project scaffolding. The architecture is a collection of standalone Node.js scripts, agent skill definitions, and HTML templates rather than a monolithic application. It relies on external services (ElevenLabs, Kie.ai, HyperFrames CLI) and CDN-hosted libraries (GSAP, Google Fonts). The repo is actively maintained with recent commits and a CI pipeline, though it exhibits some duplication between mirrored skill directories and draft-status cards that have not all been individually rendered.
This repository is a curated learning-resource aggregator (an 'awesome list') for the Zig programming language, not a software application. Classified as 'application' only because of its CI/CD tooling; it is not a data_research repo either — there are no datasets, notebooks, pipelines, or application source code. The core asset is README.md, a ~200-entry categorized collection of links to Zig books, videos, podcasts, blog posts, GitHub projects, and AI/ML frameworks, organized into ~25 categories with older entries progressively moved to a yearly archive/ directory. The only code-adjacent artifacts are package.json (markdownlint-cli, markdown-link-check), two GitHub Actions workflows, and three OpenCode agent-skill definitions.
FinClaw is an open-source, finance-domain-specific autonomous AI agent execution framework, primarily targeting Chinese A-share markets and institutional banking/insurance/trust workflows. The repository contains over 1000 modular 'skills' - standalone CLI scripts that perform market data retrieval (via AkShare, Tushare, Yahoo Finance), risk scoring (keyword-based sentiment, threshold alerts, Brinson attribution), compliance checks (sensitive phrase detection, clause conflict analysis), and report generation (Markdown/JSON outputs). Each skill is self-contained with its own entry point and dependencies, designed to be orchestrated by the FinClaw Agent OS. The repository is structured as a flat collection of independent skills under `skills/`, with significant duplication of identical scripts (e.g., `analyze.py` templates, `keyword_retrieval.py`, `pack.py`) across multiple skill directories. This copy-paste pattern reduces maintainability but allows rapid creation of domain-specific assistants without shared library dependencies. Notable well-crafted components include the cn-stock-data adapter layer (abstracting multiple Chinese data sources) and the async MCP connection factory for skill networking. The codebase is heavily biased toward Chinese financial regulations and terminology, with skills for bank loan pre-screening, fund portfolio diagnostic, insurance underwriting, and securities trading strategy backtesting. Many scripts are placeholders or demos with hardcoded positive/negative keyword lists and no real AI/ML logic, relying instead on rule-based scoring. The repository is actively maintained (last commit 2026-05-13) but suffers from medium technical drift due to duplicated boilerplate, hardcoded API tokens, and fragile web scraping patterns.
bioSkills is a curated, archived collection of over 560 bioinformatics 'skills' designed to be consumed by AI coding agents (Claude Code, Codex CLI, etc.). Each skill consists of a SKILL.md containing expert guidance, a usage-guide.md, and one or more self-contained example scripts (Python, R, shell). The repository covers the full breadth of bioinformatics—from sequence manipulation, alignment, and variant calling to single-cell, spatial transcriptomics, and causal genomics. The examples are meticulously commented with version-pinned dependencies and explicit warnings against common pitfalls, reflecting a strong pedagogical focus. The repository is not a live application or library; it is a static knowledge base that has been archived with no further updates planned. Its primary value lies in the expert-compiled recipes and anti-pattern guidance that can steer AI-generated code toward correct methodology.