everyday-skills

Plugin everyday-skills from marketplace hgkahng in hgkahng/skills. Commands read from the repo's manifest — 14 days ago.

Skills (10)

• align

  Align with the user's thoughts, intentions, and preferences before executing a task by running a focused back-and-forth question-and-answer dialogue. Use this skill whenever the user says things like 'let's align', 'align with me', 'ask me first', 'make sure you understand', 'before you start', or a

• beamer-lecture

  Create, edit, convert, debug, and visually polish academic Beamer lecture slides for Hyungu Kahng (math-dm / DIGB174; DIGB164 BP; DIGB225 ML; DDM 676 IS). Use when building or revising lecture decks, adding or rewriting frames, fixing Beamer/LaTeX errors, converting notes to slides, or tightening pl

• cut-to-the-chase

  Give the reader only what they need to understand and act on your answer. Related is not the same as relevant: leave out rejected options, dead ends, the checks you ran, side facts, and anything else from your own reasoning that the reader's understanding or next step does not depend on. Use when th

• inception

  Turn behavioral rules for AI agents into the higher-level concept they derive from. Rule lists ('do this, don't do that') invite letter-compliance and cover only anticipated cases; this skill finds the principle that makes the desired behavior the natural one, states it overtly and positively with i

• ml-research-code

  Write, review, and fix Python code for ML research: training scripts, data pipelines, utility modules, and experiment infrastructure. Use when the user says 'write code for', 'implement', 'code this up', 'build a pipeline', 'review this code', 'check my training loop', 'fix this code', 'what's wrong

• ml-research-critique

  Help design and refine ML experiments through structured dialogue. Use when the user asks to design an experiment, debug methodology, refine experimental setup, sanity check an idea, or get feedback on research direction. Triggers include 'help me design an experiment', 'I'm not sure how to test X',

• model-council

  Consult a council of AI models on the same question in parallel, then combine their answers. Three combine modes — synthesize (consensus + dissent), best-of-N judge, critique panel — and three backends: parallel subagents of the host agent (e.g. 2× Opus + 3× Sonnet), local agent CLIs (claude/codex/c

• plain-english

  Write the way an ordinary, clear speaker would say it out loud to a smart person who does not know the field: common words, short sentences, no jargon or idioms. Use when the user says 'plain English', 'explain simply', 'simpler words', 'I don't understand', 'that's hard to follow', 'use easier lang

• python-style

  Write or rewrite Python code to match the user's personal coding style preferences. Use this skill whenever the user asks to write, refactor, clean up, or review Python code for style — including requests like 'write this in my style', 'clean this up', 'rewrite this', 'does this match my preferences

• tikz-figure

  Create, iterate on, and compile publication-ready TikZ/LaTeX figures for academic papers and Beamer presentation slides. Use this skill whenever the user wants to draw, design, or refine any figure, diagram, chart, or illustration using TikZ — including neural network architectures, flowcharts, comm

Similar plugins