A Claude Code skill that evaluates whether rewriting your prompt would meaningfully improve it, then rewrites only when worth it. Triggers on 'rewrite my prompt', 'fix the typos in this prompt', 'make this clearer', or short/ambiguous prompts. Displays the rewrite and waits for approval.
Expert Prompt Engineer with Context Engineering, Meta-Prompting, Chain-of-Thought, Few-Shot, Agent Design, 50+ Template Library, and A/B Testing for creating optimal prompts and AI agents
4 multi-agent collaboration skills: fan-out planner, context budget with proposal-only memory, debate, and an acceptance gate that reconciles lane output; 3 redirect stubs remain until 0.7.0.
Free and open source, LibreJustice gives your agent ~3.9 million French-language and European decisions and ~2.1 million legal articles in force and ~1.8 million historical versions, readable as they stood on any date and linked article by article. Four MCP tools over OAuth, two research skills.
Cria carrosseis de Instagram prontos pra publicar e entrega um editor visual (Carousel Studio) ja populado: tipografia, destaque de palavra, imagem de fundo, sombra/overlay, cantos editaveis, logo e exportacao em ZIP.
Separate, optional planning skill: before non-trivial work, write an 8-line plan header that aims at the outcome above the task and names how the result will be checked.
Always-on efficiency discipline — triage token spend to task size, write minimal correct code, take fewer round-trips, keep context lean, use memory, route to the right tool. Three skills (phobos, phobos-code, phobos-plan) plus guard/state/status hooks.
Always-on efficiency discipline — triage token spend to task size, write minimal correct code, take fewer round-trips, keep context lean, use memory, route to the right tool. Three skills (phobos, phobos-code, phobos-plan) plus guard/state/status hooks.
Makes Claude write like a sharp, plain-spoken human. Loads 29 researched anti-AI-writing patterns at session start and checks every file it writes for tells. /copywrite audits or writes website copy against 36 researched copywriting rules.
Small tools that keep a repository and a work session honest without editing tracked files: doc-truth checks that what the documents say exists, handoff writes an end-of-session note whose citations are verified, todo-sweep triages TODO and FIXME comments on a local board. Python standard library and git only.
Local approve/reject (AR) board and question form: the agent proposes N options or asks N questions, you answer with the keyboard and comments, and the agent acts only on what you decided and keeps your preferences for the next rounds.