Improve an existing 10xDevs 3.0 prework lesson via an interactive decision-gate interview (which weaknesses to fix, sections to reshape, workflow change to clarify, failure modes to add, 3 takeaways to set, recommendation posture to enforce, optional /autoresearch) anchored in the fixed RC improvement report and fixed prework style guide. Use when the user asks to update, improve, polish, or apply RC report feedback to a specific lesson in src/content/lessons10xDevs3Prework. Takes the lesson markdown file path as the parameter. Must run the interview, present an improvement plan, and get explicit approval before editing.
Write or revise a Polish learner-facing 10xDevs lesson draft from the workbench lesson schema, approved lesson spec, and optional grounding brief. Use this skill whenever the user asks to draft, write, expand, continue, or reshape a main-course lesson, especially from workbench/lessons-schema.json, workbench/lessons/<lessonId>/lesson-spec.md, workbench/lessons/<lessonId>/lesson-grounding.md, or a lesson schema slot. This skill must anchor the draft in the schema, neighboring lessons, prework continuity, grounding sources, structural logic map, and workbench/references/style.md so the prose follows the 10xDevs house style without duplicating prework, stealing scope from adjacent lessons, or jumping between ideas without a clear transition.
Edit a Polish 10xDevs workbench lesson draft into release-candidate quality prose. Use this skill whenever the user asks to polish, refine, improve, make less AI-sounding, make publication-ready, edit in Polish, or prepare workbench/lessons/<lessonId>/lesson-draft.md for RC. This skill must preserve the lesson spec, grounding, factual claims, schema boundaries, and workbench/references/style.md voice while removing generic AI prose, weak argument architecture, broken section logic, Polglish, repetition, unsupported overclaims, and detached course flow.
Ground an existing 10xDevs lesson spec or draft in credible source material before drafting or revision. Use this skill whenever the user asks to research, verify, source, ground, enrich, fact-check, or strengthen a workbench lesson with external materials, popular technical sources, Hacker News/practitioner signals, known AI/ML/programming authors, official docs, papers, repos, or current tool behavior. This skill produces workbench/lessons/<lessonId>/lesson-grounding.md and updates the target lesson entry in workbench/lessons-schema.json with source metadata so lesson-draft can consume grounded claims without turning research into unsourced prose.
Review a 10xDevs workbench lesson draft as the learner who will actually read it — a working programmer learning AI-assisted development who has completed prework and the lesson's dependsOn lessons, knows some of the material and meets the rest for the first time, and is sensitive to LLM-sounding prose. Use this skill whenever the user asks to rate a draft's quality, judge whether it is ready, read it from the student's point of view, find where a learner would get confused, lost, bored, or stop trusting the text, or check that new concepts are explained without gaps and the narrative holds. This skill reconstructs the student's prior knowledge from the schema, performs a cold first-person read of workbench/lessons/<lessonId>/lesson-draft.md, and returns per-dimension scores plus a readiness verdict. It owns the felt learner experience and defers factual verification, scope, and schema continuity to lesson-rc-review.
Review a 10xDevs workbench lesson release candidate against schema, spec, grounding, draft, video scenario, style, neighboring lessons, and external sources, including a mandatory linear coherence-and-flow pass over the draft (unpaid promises, logical holes, flow interruptions, through-line integrity). Use this skill whenever the user asks for RC review, release-candidate review, production readiness, final check, review against spec, check factual grounding, verify external claims, check lesson coherence or narrative flow, or decide whether workbench/lessons/<lessonId>/lesson-draft.md is ready for production handoff. Findings must lead, ordered by severity, with concrete evidence and required fixes.
Create or refine a schema-grounded lesson specification for the 10xDevs editorial workbench. Use this skill whenever the user wants to start a lesson, pick up a lesson from lessons-schema.json, plan lesson content, define learning outcomes, explore lesson direction, diagnose flaws in a lesson idea, or prepare inputs for lesson drafting/video scenarios. This skill must run discovery first, ask decision-gate questions, surface weaknesses and alternative directions, map the lesson's internal logic before prose exists, and keep every recommendation anchored to workbench/lessons-schema.json plus the prework summary in workbench/references/prework.md so lessons do not drift away from the course journey or repeat prework unnecessarily.
Create, render, and audit mermaid diagrams for 10xDevs 3.0 lesson content. Three modes: (1) render — extract mermaid blocks from lesson markdown and render to SVG+PNG, (2) audit — validate diagrams against the style guide, (3) generate — create a new diagram from a description. Use this skill whenever the user mentions mermaid diagrams, rendering diagrams, diagram style compliance, creating a flowchart for a lesson, checking diagram consistency, or asks to visualize a concept for a lesson draft. Also trigger when the user says "render diagrams", "audit diagrams", "generate diagram", "wyrenderuj diagramy", "sprawdź diagramy", "narysuj diagram", "wygeneruj mermaid", or references the mermaid style guide.