robertnowell

Fetch engagement metrics from all platforms and show how posts are performing. Sends a Slack digest if configured.

Generate social media posts for a project. Draft them yourself following the voice rules below, then validate with the anti-slop gate.

Add a new open source project to the marketing pipeline. Researches real user pain, auto-detects project type, generates everything.

Run the daily post rotation — drafts and posts for all live projects across their channels.

Launch a project to all directories, generate drafts for all channels, and post everywhere. The full launch sequence.

Publish a draft to a social channel (Bluesky, Dev.to, Hashnode, or Mastodon).

Check credential configuration and API connectivity.

Show pipeline overview: live projects, post counts, and current state.

Deep multi-source research with source quality evaluation and hard quality gates. Launches parallel research agents, classifies sources by tier, cross-references findings, and synthesizes structured reports. Use when asked to research, investigate, or deeply analyze any topic requiring multi-source evidence.

Stages 8–9 of the video-essay pipeline. After 24h since shorts publish, pull per-angle metrics from YouTube Shorts + Instagram Reels + TikTok, compute viral scores, pick a winner. Then user-approval checkpoint on whether to promote to long-form. Use when shorts are published ≥24h ago and shorts_metrics.json is absent or stale.

Stages 0–3 of the video-essay pipeline. Preflight credentials, confirm topic with user, run broad angles research, pick 3 meaningfully different angles, write packages.md. Use this skill when starting a fresh video-essay topic OR when an episode directory has no packages.md yet. Ends at a user-approval checkpoint on the 3-angle selection.

Stage 10 of the video-essay pipeline. Expand a winning short into a 10-minute long-form video essay. Locks title + thumbnail first, then runs deep research → script → auto-critique → audio → images + QA → video assembly → publish. Use when the winner angle is set in packages.md and long-form video.mp4 is absent.

Stages 4–5 of the video-essay pipeline. For each of 3 angles — per-angle hook research, write a short script, pre-audio judge, then audio/images/covers/captioned video. Use this skill when packages.md exists but the 3 shorts aren't all assembled yet. Runs after ve-intake, before ve-publish-shorts.

Stages 6–7 of the video-essay pipeline. Visual-only Bible-compliance gate per angle, then publish all 3 shorts natively to YouTube Shorts + Instagram Reels + TikTok with ~15 min staggering. Use when all 3 shorts have video.mp4 + covers rendered but upload_result.json is absent. Ends by starting a 24h timer.

Shorts-first video essay pipeline. Three shorts enter, one wins at 24h, becomes a 10-min long-form. Use when the user says "video essay", "make a video about X", "start an episode", or invokes /video-essay. This is the dispatcher — it reads current episode state and routes to the right stage-specific sub-skill.

Analyze your Claude collaboration style and generate a skill tree visualization with character archetype card. Use when the user says "skill tree", "show my skills", "analyze my style", "check my ai fluency", or wants to see their AI fluency profile.

Deep multi-source research with source quality evaluation and hard quality gates. Launches parallel research agents, classifies sources by tier, cross-references findings, and synthesizes structured reports. Use when asked to research, investigate, or deeply analyze any topic requiring multi-source evidence.

Use when the user asks how to say something in another language, wants to express an idea in a foreign language, asks about phrasing or wording in any non-English language, or asks how to pronounce something. Trigger phrases include "how do I say", "how would a native speaker say", "what's the right way to say", "how to pronounce", "say X in Y language".

Generates a concise monthly investor update for a startup. Compiles the month from the founder's own tools (GitHub PRs, Slack activity, or custom adapters), distills it to a headline-only brief, and uses Kopi to turn that brief into an on-brand HTML email template you can send from any email platform (Klaviyo, Mailchimp, etc.). Use when the user says "investor update", "monthly update", "draft the investor update", or names a month to recap.

When the user needs a credential to make an API call, store it once via a system popup, then reference it by name in any command. The value lives only in the macOS Keychain and the broker subprocess; it never enters this conversation.

macOS sound + notification capability. Fire immediately (notify) to alert the user — useful when finishing a long task, completing a multi-step operation, or surfacing important state and the user may have stepped away. Also runs named countdown timers (start/list/cancel) for user-requested timers and pomodoros. Use proactively when ending a turn after long-running work; use on request when the user says "set a timer for X", "start a pomodoro", "what timers are running", "cancel my timer".

Purchase a domain name via the Namecheap API with a hard spending cap and explicit human confirmation. Use when the user says "buy <domain>", "register <domain>", "purchase <domain>", asks Claude to acquire a domain, or wants to check availability and price before buying. Drives the full flow inline in the conversation — Claude asks questions, runs the API calls, confirms with the user before any spend.

Bakes evaluation thinking into model output work. Every plan names what success looks like, where things can go wrong, and proposes simple checks validating key assumptions.

Analyze your Claude collaboration style and generate a skill tree visualization with character archetype card. Use when the user says "skill tree", "show my skills", "analyze my style", "check my ai fluency", or wants to see their AI fluency profile.