AvyanshKatiyar

claim-scorecard

Track public advisors' (finfluencers, analysts, newsletter writers) checkable predictions to their outcomes — extract falsifiable claims from transcripts, newsletters, or pasted text into a deduped JSONL ledger, auto-expire claims past their window, resolve outcomes via web search, and render per-advisor scorecards with hit rates. Use when the user says "score this advisor", "track these predictions", "did his calls come true", "claim scorecard for <channel>", "log these calls", "how accurate is this analyst", or pastes advisor content and wants its predictions held to their windows.

explain-my-app

Generate a personalized, plain-English explanation of the user's own codebase as a linked doc-set living in their repo, then track comprehension over time — re-explain only what changed since the last run, and quiz them from their own docs. Use when the user says "explain my app", "explain this codebase to me", "I vibe-coded this and don't understand it", "re-explain what changed", or "quiz me on my codebase".

interview-dossier

Build an adversarially-verified interview dossier on a company and one named interviewer for a specific high-stakes interview — four tiers from verified facts to a night-before decision frame, heavy web research, every claim sourced. Use when the user says "interview prep for <company>", "build a dossier", "I have a founder round", "research this company for my interview", "prep me for the CEO round", or wants serious research before a final/skip-level/founder interview rather than generic question lists.

workflow-distiller

Mine your own Claude Code history for the skills you should have. Scans ~/.claude/projects session transcripts (read-only, local-only, aggregates-only) for repeated friction→fix patterns, proposes ranked skill candidates, and drafts SKILL.md files following best practice. Use when the user asks "what skills should I make", "distill my workflows", "mine my transcripts", "turn my habits into skills", "find my repeated workflows", or wants skill ideas grounded in what they actually do every week.

create-ad

Create an OpenAI ChatGPT Ads campaign (campaign → ad group → chat_card ad) from a natural-language brief, created PAUSED with explicit confirmation before anything touches the API. Use when the user says "create an openai ad", "run a chatgpt ad", "launch a campaign on chatgpt ads", "promote X on chatgpt", or gives an ad brief mentioning OpenAI/ChatGPT ads.

insights

Analyze OpenAI ChatGPT Ads performance in natural language — spend, impressions, clicks, CTR/CPC/CPM, best and worst performers, daily trends. Use when the user asks "how are my chatgpt ads doing", "which ad performs best", "spend this week", "show CTR by campaign", or any performance question about OpenAI/ChatGPT ads.

manage

Manage OpenAI ChatGPT Ads — activate, pause, rename, rebudget, retarget, or archive campaigns, ad groups, and ads, with confirmation gates on spend-starting and irreversible actions. Use when the user says "pause my chatgpt ads", "activate the campaign", "change the budget", "update the ad copy", "archive that campaign", or any management request for OpenAI/ChatGPT ads.