Public

Session-end memory discipline. Detects which projects were touched, writes updates to WORKBENCH.md (volatile state) + skills + persistent memory, and proposes CLAUDE.md edits. Two-tier — obvious low-blast-radius writes are applied automatically and reported; evergreen/destructive changes ask first. Use when the user says "/wrap", "wrap up", "end of session", "let's close out", or when the session is clearly winding down.

Gemini vision-powered PDF to markdown converter. Handles scanned docs, multi-column layouts, tables, footnotes, flowcharts, and degraded documents that text-based extraction destroys. Uses two-tier model routing (cheap model for clean digital pages, capable model for everything else) with per-chunk confidence scoring, anti-hallucination detection, and continual learning from user corrections. Use when a user needs to convert, extract, analyze, or process any PDF document — especially scanned documents, government forms, legal contracts, academic papers, or anything where pypdf/pdfplumber returns garbage or nothing.

Fetch RSS feeds and generate AI-summarized daily digests

Scrape web pages to clean markdown with optional AI summaries

Evaluate Amazon KDP niche profitability using BSR analysis, keyword volume, and competition scoring. Use when researching book niches, validating publishing ideas, or comparing market opportunities.

Write clear, human prose. Strips AI patterns, enforces hard constraints, optionally applies a voice style. Use when writing or editing any prose, when asked to "make it sound human," "remove AI tells," "write clearly," or when producing any content that must not read as AI-generated. Load style references for specific voices (Cowen, Koe, Levine-Berry, Lewis, Pirate Wires, OpenEd, Charlie) or social formats (Trung Phan).

Run Gemini Deep Research from Claude Code. Optimizes prompts for depth and structure, then executes multi-step research via the Interactions API. Use when you need in-depth analysis, competitive landscaping, literature reviews, market research, or any question requiring real-time web research across many sources.

Delegate large-context writing tasks to Gemini's 1M token window. Send an entire manuscript + detailed instructions in a single pass. Use for book rewrites, structural refactoring, redundancy analysis, and any writing task where the AI needs to see everything at once.

Export markdown files to Notion with proper rich text formatting. Use when staging articles, drafts, or documentation in Notion. Handles bold, italic, links, headers, bullets, numbered lists, and code blocks.

Find, install, create, improve, and publish AI agent skills through the Sundial ecosystem. Use when the user wants to find or search for skills, install a skill, create a new skill, improve or evaluate an existing skill, or publish a skill to Sundial Hub. Trigger phrases include "find a skill", "install skill", "create a skill", "make a skill", "improve this skill", "evaluate skill", "publish skill", "push skill", "search for skills".

Write long-form essays and newsletters in Dan Koe's voice — philosophical depth made accessible, staccato rhythm with expansive passages, confident authority, zero hedging. Structured as a teaching conversation with bad-AI/correction/good-version rounds reverse-engineered from his actual articles. Use for newsletters, X articles, blog essays, or any long-form content that needs to blend philosophy with practical frameworks.

Transcribe podcast episodes from RSS feeds using Gemini Flash

Transcribe YouTube videos and playlists using Gemini Flash

Use this skill whenever the user wants to fill out a PDF form, analyze PDF form fields, troubleshoot form field mapping, verify field identification accuracy, or export a filled PDF. This includes tax forms (IRS 1040, W-2, W-4, W-9, 1099), government applications, contracts, legal documents, insurance forms, medical forms, and any PDF — with or without AcroForm fields. Also triggers when the user reports wrong field mapping ("it put 1f into 1d"), vision analysis failures, wants to improve field extraction accuracy, or needs to fill a flat/non-fillable PDF via annotations. If the user mentions a fillable PDF, asks to populate form fields, or wants to write text onto a flat PDF form, use this skill.

Extract actionable Claude Code skills from raw source material — transcripts, conversations, workflows, expertise dumps. This skill identifies repeatable, promptable workflows embedded in content and scores them by leverage. Use when processing a corpus (podcast transcripts, blog posts, course material) to discover what skills could be built from it.

Generate AI videos using Google VEO 3.1 or OpenAI Sora. Two providers for different strengths - VEO for native audio, Sora for visual quality and longer clips.

Generate tweets and threads in the style of Trung Phan. Not just voice — captures his humor mechanics, format taxonomy, topic selection filter, and structural patterns. Use for trend-reactive tweets, meme commentary, and business/culture threads.

Write in Tyler Cowen's style - matter-of-fact, understated, treats enormous ideas as obvious observations. Read the passages. Absorb the flatness. Channel the HOW, not the content.

Write in a combined Matt Levine + Wendell Berry voice. Levine's dry logic-walking and parenthetical humor for the analytical sections. Berry's meditative patience for the human ones. Read the passages. Absorb the rhythm. Channel the HOW, not the content.

Rewrite nonfiction prose in the style of Michael Lewis. Strips formula, scaffolding, and manufactured drama. Replaces them with earned narrative, invisible transitions, and evidence that draws its own conclusions.

Interpret blood work through Ray Peat's metabolic framework -- thread-based analysis connecting markers into metabolic stories. Paste your labs and get the story your body is telling.

Generate AI images using Gemini image generation API. Use this skill when content needs images - thumbnails, social posts, blog headers, or creative visuals. Follows an iterative workflow - brainstorm concepts, select direction, generate in multiple styles, then produce via API.