This repository is a local-first, source-cited knowledge base management system, delivered primarily as a Claude Code plugin and a standalone Python CLI tool. It manages Obsidian vaults with a strong emphasis on data integrity, provenance tracking, and deterministic operations. The architecture is a monolithic Python library (`claude_obsidian/`) with a formal transaction system for all vault mutations, a robust lint engine, and a retrieval pipeline (BM25 + optional semantic reranking). It features a sophisticated skill/agent framework for AI integration across multiple hosts (Claude, Cursor, Windsurf, etc.). While primarily an application, it incorporates significant data engineering and research data management (RDM) patterns. The provenance ledger system (`ledgers.py`) enforces strict schemas for sources and claims, including canonical URL handling, content-addressed identity, and evidence validation rules. The retrieval pipeline implements Anthropic's Contextual Retrieval pattern. This hybrid nature makes it a powerful tool for both knowledge management and structured research workflows. The codebase demonstrates exceptionally high craftsmanship in its transaction system, defensive I/O, and deterministic analysis tools. It is actively maintained with a comprehensive test suite and a formal release process. Its primary limitation is platform dependency on Linux/macOS for write operations (WSL required for Windows).
CCteam-creator is a Claude Code skill that enables multi-agent team orchestration through file-based planning and role-based collaboration. It sets up persistent agent teams with specialized roles (backend-dev, frontend-dev, researcher, e2e-tester, reviewer, custodian) that communicate via SendMessage while maintaining state in .plans/ directories. The system includes mechanisms for context recovery, team snapshots, doc-code synchronization, invariant-driven review, and automated code quality checks via golden_rules.py. Team-lead (the main conversation) acts as the control plane managing task delegation, phase gates, and operational protocols.
This repository is a content-based Claude Code skill package rather than a conventional software service. It contains two agent skills — /tutor-setup and /tutor — that an LLM coding assistant executes using filesystem and WebFetch tools. tutor-setup scans a working directory, auto-detects whether it contains source code or documents, and generates an Obsidian StudyVault: concept notes, dashboards, exam-traps, and practice questions, with strict rules for source traceability, tag registries, wiki-linking, and folder conventions. tutor-setup operates through detailed phase models (D1-D9 for documents, C1-C9 for codebases), and codebase mode produces architecture maps, module notes, API surfaces, and onboarding exercises. tutor then reads that vault and runs interactive 4-question quiz rounds, updating concept-level progress files and a compact dashboard; it supports diagnostic, weak-area drill, section review, and hard-mode sessions, all governed by a zero-hint quiz design rule in references/quiz-rules.md. The implementation is almost entirely Markdown instructions plus two small Bash install/uninstall scripts; there is no compiled code, test suite, CI pipeline, or dependency manifest. The operational logic is prompt/instruction engineering that Claude Code interprets. External dependencies are Claude Code (or a compatible npx-skills-aware assistant), Obsidian for reviewing generated vaults, pdftotext/poppler for PDF extraction, and the assistant's built-in Read, Write, Edit, Glob, Grep, Bash, WebFetch, and AskUserQuestion tools. The data flow is unidirectional: source files in the CWD are scanned and extracted, transformed into StudyVault markdown, then quiz answers update concept files and dashboard aggregate stats. The skill enforces a CWD boundary to avoid accessing files outside the working directory. The strong point is unusually detailed and internally consistent procedural specification: mandatory source-content verification, per-topic coverage rules, tag registries, active-recall fold callouts, interlinking requirements, and self-review quality checklists. Weaknesses are the absence of automated tests/CI and versioned dependency pins, and behavior ultimately depends on the host LLM following prompt rules correctly. The README also references a different GitHub owner (RoundTable02) than the repository path (bevibing/tutor-skills), indicating a fork or repackage whose provenance should be verified before use. Overall, it fulfills its stated purpose well, and several Markdown modules are genuinely reusable as templates for other agent skills, study-vault generators, or adaptive quiz systems.
This repository provides a Claude Code skill for literate programming, along with supporting CLI tools. The skill (SKILL.md) defines a slash command that transforms a codebase into a `.lit.md` file – a single Markdown document that weaves prose, diagrams, and named code chunks together. The actual tools are three TypeScript scripts: `tangle.ts` extracts source files from a `.lit.md`, `untangle.ts` reverse-syncs edits back into the `.lit.md`, and `hook-reverse-sync.ts` is a Claude Code PostToolUse hook for automatic reverse-sync. The architecture is simple: standalone scripts with no external dependencies beyond Bun's runtime. The tangler parses fenced code blocks with `{chunk="name" file="path"}` attributes, builds a dictionary of chunks (supporting additive chunks), and recursively expands `<<reference>>` with cycle detection and indent preservation. The untangler uses a `.lit.map.json` source map to splice edited lines back into the original `.lit.md`, then re-tangles. The code is clean, well-documented, and demonstrates good engineering practices for a small tool. It is not a data research project nor a typical application – it is a developer tool for literate programming workflows.
This repository is a Claude Code skill that enables an AI agent to audit bank transaction CSVs, identify recurring subscriptions, and automate cancellations via browser control. It contains no traditional application code; instead, it provides a structured instruction set (SKILL.md), an interactive HTML report template (template.html), a reference guide for cancellation URLs and dark patterns (common-services.md), and a JSON dataset of common subscription pricing (subscriptions.json). The skill is designed to be used with Anthropic's Claude Code CLI, which executes the workflow: reading CSV files, asking the user to categorize subscriptions, generating a local HTML audit report, and optionally using browser automation to cancel services. The repository is well-documented with a getting-started guide and a Vercel-hosted landing page, but it is not a standalone application—it is a configuration/prompt package for an AI agent.