The Playbook

The durable rules of engagement: every Claude Code skill that's installed and every behavioral rule Claude follows. One tap on any card opens a prefilled Telegram message to tweak or remove it.

Skills

55 installed

adapt

2.1.1 user-invocable

Adapt designs to work across different screen sizes, devices, contexts, or platforms. Implements breakpoints, fluid layouts, and touch targets. Use when the user mentions responsive design, mobile layouts, breakpoints, viewport adaptation,

ai-morning-brief

Generate Dan's daily AI/tech morning briefing โ€” Claude Code tips, OpenClaw updates, AI news, and practical angles for HMG and Roundup app. Sends to Telegram. Use when Dan wants his morning AI briefing or calls /ai-morning-brief.

animate

2.1.1 user-invocable

Review a feature and enhance it with purposeful animations, micro-interactions, and motion effects that improve usability and delight. Use when the user mentions adding animation, transitions, micro-interactions, motion design, hover effect

audit

2.1.1 user-invocable

Run technical quality checks across accessibility, performance, theming, responsive design, and anti-patterns. Generates a scored report with P0-P3 severity ratings and actionable plan. Use when the user wants an accessibility check, perfor

awesome-design-md

Apply a real-world design system to a project. Use when the user says "make it look like X", "use Stripe's design", "Airbnb-style UI", or wants pixel-perfect UI matching a known brand. Drops a DESIGN.md into the project root so Claude gener

bcg-prospect-matrix

Classify HMG prospects into BCG Matrix quadrants (Star/CashCow/QuestionMark/Dog) to guide enrichment and draft queue priorities. Use when reviewing the pipeline, deciding which prospects to enrich next, or archiving low-value records.

bolder

2.1.1 user-invocable

Amplify safe or boring designs to make them more visually interesting and stimulating. Increases impact while maintaining usability. Use when the user says the design looks bland, generic, too safe, lacks personality, or wants more visual i

brw-ai-discoverability-audit

Audit how HMG (or any brand) appears in AI-powered search systems โ€” ChatGPT, Perplexity, Claude, Gemini. Use when Dan asks about HMG's AI presence, how to show up in ChatGPT for queries like "treasury management for PE sponsors", "Schwab cu

brw-newsletter

Create B2B newsletters for HMG โ€” industry-specific, role-aware, geography-aware. Use when Dan wants to draft a newsletter, nurture sequence, or thought leadership email for HMG's RE sponsor audience.

business-development

HMG partnership outreach, market research, competitor analysis, and proposal generation. Use when identifying potential HMG referral partners (RE/PE-focused law firms, fund admins, accountants serving sponsors, brokerages, specialty lenders

clarify

2.1.1 user-invocable

Improve unclear UX copy, error messages, microcopy, labels, and instructions to make interfaces easier to understand. Use when the user mentions confusing text, unclear labels, bad error messages, hard-to-follow instructions, or wanting bet

cold-outreach

Generate high-converting cold outreach emails, follow-up sequences, and LinkedIn messages for HMG using Hormozi-style frameworks. Use when drafting B2B cold emails, writing multi-touch sequences, evaluating outreach copy, or personalizing a

colorize

2.1.1 user-invocable

Add strategic color to features that are too monochromatic or lack visual interest, making interfaces more engaging and expressive. Use when the user mentions the design looking gray, dull, lacking warmth, needing more color, or wanting a m

competitor-analysis

Generate structured competitive analysis reports for HMG โ€” feature comparisons, pricing, SWOT, and strategic positioning vs other treasury management providers serving private capital sponsors. Use when Dan wants to understand the competiti

contract-generator

Generate professional freelance contracts, SOWs, NDAs, and retainer agreements for HMG client engagements. Use when Dan needs to send an engagement letter, statement of work, NDA, or retainer agreement to a new client.

critique

2.1.1 user-invocable

Evaluate design from a UX perspective, assessing visual hierarchy, information architecture, emotional resonance, cognitive load, and overall quality with quantitative scoring, persona-based testing, automated anti-pattern detection, and ac

dedup-outreach

Scan HMG outreach drafts and prospect pipeline for duplicate contacts or companies, consolidate duplicates, and report. Use when outreach drafts are being reviewed, before approving any batch, or when a new discovery run has completed.

defuddle

Extract clean markdown content from web pages using Defuddle CLI, removing clutter and navigation to save tokens. Use instead of WebFetch when the user provides a URL to read or analyze, for online documentation, articles, blog posts, or an

delight

2.1.1 user-invocable

Add moments of joy, personality, and unexpected touches that make interfaces memorable and enjoyable to use. Elevates functional to delightful. Use when the user asks to add polish, personality, animations, micro-interactions, delight, or m

ckm:design-system

Token architecture, component specifications, and slide generation. Three-layer tokens (primitiveโ†’semanticโ†’component), CSS variables, spacing/typography scales, component specs, strategic slide creation. Use for design tokens, systematic de

distill

2.1.1 user-invocable

Strip designs to their essence by removing unnecessary complexity. Great design is simple, powerful, and clean. Use when the user asks to simplify, declutter, reduce noise, remove elements, or make a UI cleaner and more focused.

harden

2.1.1 user-invocable

Make interfaces production-ready: error handling, empty states, onboarding flows, i18n, text overflow, and edge case management. Use when the user asks to harden, make production-ready, handle edge cases, add error states, design empty stat

hilltopmanagementgroup-design

Design system skill for hilltopmanagementgroup. Activate when building UI components, pages, or any visual elements. Provides exact color tokens, typography scale, spacing grid, component patterns, and craft rules. Read references/DESIGN.md

hmg-brief

Generate a full HMG prospect brief for a given firm name โ€” research, ICP fit score, recommended outreach angle, and artifact plan. Usage: /hmg-brief <firm name>

hmg-deep-research

Deep multi-step research on any HMG topic โ€” prospect firms, market intelligence, fund structures, competitive landscape, or RE market trends. Use when a research question requires more than a single query, needs synthesis across multiple so

hmg-market-brief

Generate a daily or weekly HMG market intelligence brief covering RE finance news, fund formation activity, and market conditions relevant to RE sponsors. Use when Dan asks for a market update, daily brief, or "what's happening in RE financ

hmg-pipeline-report

Generate a live HMG pipeline status report โ€” prospect counts by tier, outreach drafts pending, sessions active, and key gaps. Use when Dan asks for a pipeline update, "how many prospects do we have," or "what's the status of outreach."

hmg-prospect-lookup

Research and enrich a specific RE sponsor firm or contact using Tavily web search. Use when given a firm name or contact to research before outreach, when enriching a prospect that has low completeness score, or when Dan asks "who is [firm/

huashu-design

่Šฑๅ”Design๏ผˆHuashu-Design๏ผ‰โ€”โ€”็”จHTMLๅš้ซ˜ไฟ็œŸๅŽŸๅž‹ใ€ไบคไบ’Demoใ€ๅนป็ฏ็‰‡ใ€ๅŠจ็”ปใ€่ฎพ่ฎกๅ˜ไฝ“ๆŽข็ดข+่ฎพ่ฎกๆ–นๅ‘้กพ้—ฎ+ไธ“ๅฎถ่ฏ„ๅฎก็š„ไธ€ไฝ“ๅŒ–่ฎพ่ฎก่ƒฝๅŠ›ใ€‚HTMLๆ˜ฏๅทฅๅ…ทไธๆ˜ฏๅช’ไป‹๏ผŒๆ นๆฎไปปๅŠกembodyไธๅŒไธ“ๅฎถ๏ผˆUX่ฎพ่ฎกๅธˆ/ๅŠจ็”ปๅธˆ/ๅนป็ฏ็‰‡่ฎพ่ฎกๅธˆ/ๅŽŸๅž‹ๅธˆ๏ผ‰๏ผŒ้ฟๅ…web design tropesใ€‚่งฆๅ‘่ฏ๏ผšๅšๅŽŸๅž‹ใ€่ฎพ่ฎกDemoใ€ไบคไบ’ๅŽŸๅž‹ใ€HTMLๆผ”็คบใ€ๅŠจ็”ปDemoใ€่ฎพ่ฎกๅ˜ไฝ“ใ€hi-fi่ฎพ่ฎกใ€UI mockupใ€prototypeใ€่ฎพ่ฎกๆŽข็ดขใ€ๅšไธชHTML้กต้ขใ€ๅšไธชๅฏ่ง†ๅŒ–ใ€appๅŽŸๅž‹ใ€

impeccable

3.0.4 user-invocable

Use when the user wants to design, redesign, shape, critique, audit, polish, clarify, distill, harden, optimize, adapt, animate, colorize, extract, or otherwise improve a frontend interface. Covers websites, landing pages, dashboards, produ

interface-design

Craft-first interface design for dashboards, admin panels, SaaS apps, tools, settings pages, data interfaces, and interactive products. Use when designing, building, reviewing, auditing, or refining product UI where visual craft, layout hie

layout

2.1.1 user-invocable

Improve layout, spacing, and visual rhythm. Fixes monotonous grids, inconsistent spacing, and weak visual hierarchy. Use when the user mentions layout feeling off, spacing issues, visual hierarchy, crowded UI, alignment problems, or wanting

obsidian-cli

Interact with Obsidian vaults using the Obsidian CLI to read, create, search, and manage notes, tasks, properties, and more. Also supports plugin and theme development with commands to reload plugins, run JavaScript, capture errors, take sc

obsidian-markdown

Create and edit Obsidian Flavored Markdown with wikilinks, embeds, callouts, properties, and other Obsidian-specific syntax. Use when working with .md files in Obsidian, or when the user mentions wikilinks, callouts, frontmatter, tags, embe

optimize

2.1.1 user-invocable

Diagnoses and fixes UI performance across loading speed, rendering, animations, images, and bundle size. Use when the user mentions slow, laggy, janky, performance, bundle size, load time, or wants a faster, smoother experience.

overdrive

2.1.1 user-invocable

Pushes interfaces past conventional limits with technically ambitious implementations โ€” shaders, spring physics, scroll-driven reveals, 60fps animations. Use when the user wants to wow, impress, go all-out, or make something that feels extr

polish

2.1.1 user-invocable

Performs a final quality pass fixing alignment, spacing, consistency, and micro-detail issues before shipping. Use when the user mentions polish, finishing touches, pre-launch review, something looks off, or wants to go from good to great.

quieter

2.1.1 user-invocable

Tones down visually aggressive or overstimulating designs, reducing intensity while preserving quality. Use when the user mentions too bold, too loud, overwhelming, aggressive, garish, or wants a calmer, more refined aesthetic.

sentiment-priority-scorer

Score and rank HMG prospect leads by sentiment, urgency, intent, and recency to produce a P1/P2/P3 priority tier. Use when prioritizing which prospects to draft outreach for, ranking a batch of prospects, or reviewing pipeline quality.

shape

2.1.1 user-invocable

Plan the UX and UI for a feature before writing code. Runs a structured discovery interview, then produces a design brief that guides implementation. Use during the planning phase to establish design direction, constraints, and strategy bef

skillui

Reverse-engineer any website's design system into a Claude-ready skill. Use when the user says "make it look like [site]", "extract the design from [URL]", "scan [site] for its design system", or wants to match a specific site's visual styl

ckm:slides

Create strategic HTML presentations with Chart.js, design tokens, responsive layouts, copywriting formulas, and contextual slide strategies.

typeset

2.1.1 user-invocable

Improves typography by fixing font choices, hierarchy, sizing, weight, and readability so text feels intentional. Use when the user mentions fonts, type, readability, text hierarchy, sizing looks off, or wants more polished, intentional typ

ui-improve

Full-stack UI analysis and improvement workflow. Use when the user says "analyze the UI", "make the UI better", "improve the design", "review the frontend", or similar. Orchestrates all design skills in sequence: extract โ†’ audit โ†’ critique

ckm:ui-styling

Create beautiful, accessible user interfaces with shadcn/ui components (built on Radix UI + Tailwind), Tailwind CSS utility-first styling, and canvas-based visual designs. Use when building user interfaces, implementing design systems, crea

watch

user-invocable

Watch a video (URL or local path). Downloads with yt-dlp, extracts auto-scaled frames with ffmpeg, pulls the transcript from captions (or Whisper API fallback), and hands the result to Claude so it can answer questions about what's in the v

web-design-guidelines

Review UI code for Web Interface Guidelines compliance. Use when asked to "review my UI", "check accessibility", "audit design", "review UX", or "check my site against best practices".

Behavioral rules

32 active

10-am-hmg-send-cron-canceled-indefinitely-2026-05-12

feedback-10am-send-cron-canceled.md

"The 10 AM Mon-Fri HMG send_daily_batch.py cron is canceled indefinitely by Dan as of 2026-05-12. Do not re-enable, do not route around it with a different scheduled send, do not run send_daily_batch.py ad-hoc. Outbound sends are ๐Ÿ”ด-lane unt

14-day recipient-history gate is a first-touch gate, not a follow-up gate

feedback-14d-gate-first-touch-only.md

Don't add `recent_recipients_14d()` filter to sequence_follow_up.py โ€” follow-up touches by definition target recently-contacted addresses

feedback-be-efficient-by-default

feedback-be-efficient-by-default.md

"Default to the minimum tokens/tool-calls/context that fully does the job, then stop; efficiency is the path, not a cap on rigor"

Build end-to-end meta-canary for any pipeline, not just per-step canaries

feedback-build-meta-canary-not-just-step-canaries.md

When debugging pipeline silent failures, the fix is not a better per-step alert โ€” it's an end-to-end meta-canary that catches upstream-quality issues, forward-looking starvation, and silent post-step degradation

Run Codex review on touch-generating scripts before shipping

feedback-codex-review-touch-scripts.md

When patching batch_first_touch.py or sequence_follow_up.py, propose `codex review --uncommitted` before declaring the fix shipped โ€” Codex catches bugs that grep + targeted reads miss

feedback-daily-updates-into-morning-pdf

feedback-daily-updates-into-morning-pdf.md

"When Dan asks for a recurring/daily update, fold it into the 8 AM morning PDF brief โ€” never add a new separate Telegram cron notification."

HMG prospect search uses DuckDuckGo, never Brave

feedback-ddg-not-brave-for-hmg-search.md

Per the marketing-plans/prospect-stack.md doctrine, DDG is the documented free search backend; Brave should never be added even as optional fallback

Default to Telegram reply, regardless of input wrapper

feedback-default-to-telegram-reply.md

When answering Dan, ALWAYS use the Telegram reply tool โ€” even if his message arrives without a `<channel source="plugin:telegram:telegram">` wrapper. He reads on Telegram primarily; Claude Code transcript output is invisible to him.

design-bar-client-ready

feedback-design-bar-client-ready.md

Every design request from Dan must produce client-facing-ready, outstanding output โ€” not first drafts. Run the full ui-improve pipeline (build โ†’ audit โ†’ refine โ†’ verify โ†’ polish) every time, no shortcuts.

Execute without permission, with one consolidated outcome

feedback-execute-without-permission.md

When Dan authorizes a task ("go" / "execute" / "do this for me" / "yes do it"), run the full plan with available tools (browser, API, CLI), no per-step approval gates, no progress pings โ€” ship one consolidated outcome message at the end. Th

feedback-execution-power-ladder

feedback-execution-power-ladder.md

"Default execution behavior โ€” audit before declaring done, match effort to task, escalate power when warranted, self-pace iterative work"

HMG positioning is treasury management, NOT fractional CFO

feedback-hmg-positioning.md

HMG = Hilltop Management Group, a treasury management firm for PE/RE sponsors. Never frame HMG as fractional CFO, accounting firm, or generic advisory. Anchor every HMG-related reference to "treasury management for private capital."

Don't run Codex calls without explicit approval

feedback-no-codex-without-approval.md

Codex CLI uses Dan's ChatGPT subscription auth tokens stored in ~/.codex/auth.json. Every codex command consumes his rate-limited ChatGPT quota โ€” same bucket as his web/desktop app sessions. Do not run codex exec / codex / codex review / et

No hands or faces in AI-generated marketing imagery

feedback-no-hands-faces-in-ai-imagery.md

Default photo prompts to "no people, no hands, no faces" โ€” Pollinations FLUX (and most free image-gen) botches anatomy in ways that read as AI-uncanny on Instagram/LinkedIn carousels.

feedback-no-outbound-emails-at-all

feedback-no-outbound-emails-at-all.md

"HARD STOP โ€” Claude must never send any email in any form; drafts only, Dan sends"

feedback-no-side-effecting-poll-loops

feedback-no-side-effecting-poll-loops.md

Never poll an endpoint that has user-visible side effects (Telegrams, emails, posts, charges) in a retry/until loop โ€” test deploy status via a side-effect-free path

no-stitch

feedback-no-stitch.md

"Don't propose or reach for Google Stitch. Dan evaluated it 2026-05-20 on an HMG redesign and decided his existing direction beats it."

Default to parallel sub-agent fan-out for multi-stream questions

feedback-parallel-subagent-default.md

Multi-stream tasks (status sweeps, competitor pulses, enrichment passes) launch agents in parallel, not sequential

portable-paths-in-codex-bound-messages

feedback-portable-paths-for-codex-messages.md

When drafting messages Dan will paste to Codex (on Mac), only use machine-portable paths โ€” never WSL-specific paths or Claude-Code-specific auto-memory paths

Repo visibility default โ€” PRIVATE

feedback-repo-visibility-default-private.md

When creating GitHub repos on Dan's behalf, default to private. Only make public if functionality requires it (and ask first).

Save durable facts to Obsidian in real time, not at end of day

feedback-save-durable-facts-immediately.md

When a durable fact appears (firm / contact / strategy / architecture / framework), write it to the Obsidian Brain immediately โ€” not as a catch-up at end of session. Apply memory-routing.md question #2 at the moment of learning, not deferre

/scope command โ€” quick / campaign / deep planning depth

feedback-scope-selector.md

When Dan prefixes a Telegram message with /scope quick|campaign|deep, override the default planning depth for that task โ€” quick = no brainstorm/no plan, campaign = brainstorm + plan + 2-3 parallel subagents, deep = brainstorm + research + p

Screenshot after every step in multi-step browser/CLI tasks

feedback-screenshot-per-step.md

Verify state after each step in a multi-step task, not just at the end

Self-healing โ€” fix errors before reporting them

feedback-self-healing-errors.md

When a script/cron/deploy/build fails during a task, attempt root-cause diagnosis and a fix within the autonomous lane before surfacing the error to Dan

Self-improving โ€” propose skills for recurring patterns

feedback-self-improving-skills.md

When the same multi-step task pattern recurs 3+ times or Dan asks for it twice, invoke skill-creator and propose a new skill so future runs trigger automatically

feedback-stay-on-flat-subscription

feedback-stay-on-flat-subscription.md

"Keep everything on Dan's flat Max 20x subscription; don't propose or build anything that bills separately (metered API, platform fees)"

don-t-bold-wrap-urls-in-telegram-replies-url

feedback-telegram-no-markdown-around-urls.md

"Never wrap a bare URL in markdown asterisks (`**https://...**`) in a Telegram reply unless the format is markdownv2. In plain-text mode (the default), Telegram leaves the asterisks visible and many clients include them in the tap-target โ€”

Telegram replies use status-emoji prefix

feedback-telegram-status-emoji.md

Every reply Claude sends Dan via Telegram leads with one of ๐Ÿšจ๐Ÿ”ด๐ŸŸกโœ…๐Ÿ“Š๐Ÿค”โธ so triage is fast

Default to transient HTML viz for data questions

feedback-transient-viz-default.md

Ad-hoc data questions return a clickable HTML viz attached to the Telegram reply, not a bullet list

verify-before-quoting-state

feedback-verify-before-quoting-state.md

Before quoting any machine-checkable state from BLOCKERS.md, PRIMER.md, or any memory file, run the cheap verification check first. Memory files lag reality when Dan completes work outside Claude sessions.

feedback-verify-rendered-artifacts-not-just-filter

feedback-verify-rendered-artifacts-not-just-filter.md

When declaring outbound drafts "ready to send", verify the linked artifacts (microsites, attachments, embedded URLs) are branded for the correct prospect โ€” not just that the safety filter passes

Use /watch skill for every video URL Dan sends

feedback-watch-skill-for-videos.md

When Dan shares a video URL, invoke the watch skill โ€” never fall back to manual yt-dlp + grep-VTT

Generated ยท Sat Jun 20, 2026 โ€” 10:23 PM CT 55 skills ยท 32 rules ยท tap to tweak via @dansclaude_bot