Recently updated open-source projects

Crawlora Agent Skills teach coding agents to fetch structured public data through Crawlora’s REST API, returning JSON for search, marketplaces, YouTube, app stores, reviews, finance, maps, and keyword research. The bundle includes standalone skills, endpoint references, and a small curl helper, so no MCP setup is required. Users supply a separate CRAWLORA_API_KEY, and the repository’s role is to package the instructions and request helpers rather than store the data itself.

getdocs turns a documentation site into a local Markdown tree that coding agents can search without repeatedly loading navigation, banners, or rendered HTML. It preserves URL structure, page metadata, and a crawl manifest, and prefers cloning an open-source repository’s original Markdown before falling back to crawling. Runs can target one section or a whole site, resume from output files, and report whether the result was crawled or cloned.

Rethink adds a structured second pass after an agent has proposed a plan. It audits load-bearing assumptions, checks questions against sources, considers the strongest opposing case, runs a pre-mortem, and records the delta between the original and revised plan. Use it only when a concrete plan exists and the thinking budget is raised; extra self-review is not presented as a guarantee of improvement.

Cursor Agent Tools provides a Python framework for coding assistance with Claude, OpenAI, or locally hosted Ollama models. Agents can edit and analyze code, search projects, run approved terminal commands, inspect images, and call registered functions while a permission layer controls risky operations. Cloud providers require credentials, and local tool support depends on the selected model.

agency-os turns Notion into a visual control board where ideas are discussed, scoped into dependent tasks, approved by an operator, and then assigned to agents that return result links to the same rows. Nothing runs autonomously: only explicitly marked work is dispatched, making it suitable for founders and teams that want visible status and human control; setup requires a Notion integration and an MCP-capable host or Claude Code plugin.

Pomosh keeps a Pomodoro session visible in the macOS menu bar so you can follow work and break cycles without reopening the app. Set the duration and number of cycles, then use notifications, sounds, the countdown, or a global shortcut to track progress. You can pause the timer and skip or restart the current round when plans change.

Give this skill a goal, brief, transcript, or existing backlog, and it builds a user story map around the project’s current context. It interviews for missing decisions, slices work into a feasible first release, prioritizes stories with WSJF, RICE, or MoSCoW, and can add acceptance criteria and an end-to-end test contract. It is designed for Claude Code and other hosts that support the Agent Skills standard.

vault-crystallize has two separate workflows: crystallize updates recovery state and durable knowledge in Handoff.md only when something changed, while distill converts a concrete source into traced evidence entries. AGENTS.md owns workspace and write boundaries, and the vault protocols define schemas and contradiction handling. It is deliberately not a generic document-cleanup, deep-review, memory-sync, or polished-summary tool.

antipatterns reviews drafts for stock vocabulary, rhythmic tics, hedging, vague claims, and other signals that make prose feel machine-written. It rewrites the affected passages and checks for voice drift so cleanup does not turn the piece into generic corporate copy. A configurable catalog and Python density scanner support repeatable checks, though the scanner only catches lexical patterns and leaves structural judgment to the model.

Plan a physical B2B SaaS outreach campaign from targeting through packaging. The skill researches 20–50 prospects, creates a branded PDF for each, groups stops into driving clusters, and exports Google Maps routes plus a My Maps KML; its template writes are limited to a development or internal tenant.

keepwright detects a repository's stack and installs a constitution, structured rules, GitHub Actions, portable validators, and hooks that keep architecture consistent over time. Its setup, audit, review, and overhaul workflows can derive project-specific design and writing patterns, while code and configuration changes remain human-gated. AI review workflows use CLAUDE_CODE_OAUTH_TOKEN, with an API key documented as a fallback.

Grite stores issues and task history in Git, so they branch, merge, and sync alongside the code through ordinary fetch and push operations. Agents can claim work with distributed locks, record progress and architectural memory across sessions, and model dependencies between tasks. Its deterministic merge model is designed to preserve concurrent updates without requiring an external tracker, account, database, or network service beyond Git itself.

This Claude Code skill writes and revises Brazilian Portuguese for articles, posts, scripts, newsletters, emails, sales pages, proposals, and documentation. Its references target translation-like phrasing, filler vocabulary, and repetitive structures, followed by a post-writing checklist. The stated goal is clearer, more specific prose rather than text shaped to satisfy unreliable AI detectors.

This Claude skill applies Jeremy Miner’s NEPQ method to three sales tasks: attraction posts, replies to interested leads, and live call scripts. Each mode follows questions about problem awareness, solution awareness, consequences, and commitment instead of leading with a pitch. The public skill is generic, so you provide your name, niche, service, and preferred tone before drafting.

The Programmatic SEO Expert helps design scalable page systems with unique value for both Google Search and AI retrieval. It first tests whether programmatic SEO fits the business, then covers data foundations, server-rendered HTML, answer-first structure, index management, and organic plus AI visibility measurement. A staged rollout discourages launching huge page counts before quality is proven. References also cover GEO, AEO, LLMO, crawler access, and rendering checks, while fast-changing mechanics should be re-verified against current guidance.

saasify audits an existing app, confirms scope, and adds missing SaaS foundations in dependency order. It covers authentication, organizations and database-enforced RLS isolation, RBAC, plans and billing, invites, and onboarding. The skill works with Claude Code, Cursor, Codex, and other coding agents, with adaptable templates and checks for Stripe, HYP, license, or credit-based billing.

Journalism Skill helps journalists and researchers organize investigations around retrievable sources, careful causal language, and explicit verification. Its guidance covers public-records work, interview planning, data review, sensitive allegations, and right-of-reply handling, with reusable templates and examples for both agentic and web-chat settings. It structures the work but cannot turn unsupported claims into verified facts.

Quantitative Researcher supports audit-and-reproduce, analyze, critique, and consult modes for research data and papers. It inspects coding and assumptions, flags common errors such as suppression or causal overreach, and performs calculations with Python statistics libraries before writing APA-structured methods, results, and limitations. The skill depends on an execution-capable environment; it is not designed to replace computation with model reasoning.