premdevai
- 19 skills
- 0 followers
- 6 hours ago last updated
- ▌ Pre Mortem · premdevaiA structured failure-imagination tool. The user describes a plan, project, launch, decision, or initiative; this skill assumes it's six months (or other relevant horizon) in the future and the project has FAILED, then enumerates the most likely concrete reasons why, ranked by probability and impact. Use this skill whenever the user says "do a pre-mortem", "what could go wrong", "how might this fail", "imagine this fails — why", "stress-test my plan", "what are the failure modes", or shares a plan and asks for risk analysis. Trigger on phrases like "what could kill this", "where does this break", "where are the risks". Different from devils-advocate (which argues the plan shouldn't happen) — this skill assumes the plan happens and finds where it breaks. Do not use for code-level risk review (brutal-code-reviewer) or generic critique (use roast-mode or devils-advocate).
- ▌ Roast Mode · premdevaiA savage, unfiltered roaster that tears into ideas, projects, plans, pitches, designs, resumes, tweets, life choices, hot takes, side hustles, startup ideas, and anything else the user willingly throws into the fire. Use this skill whenever the user says "roast me", "roast this", "be brutal", "destroy this", "tear it apart", "no mercy", "harsh take", "rip into this", "be savage", "give me your worst", or otherwise explicitly invites a hostile-comedic critique. Trigger even when the request is short and casual ("roast my Twitter bio", "roast this idea") — the invitation is the trigger. Do not use for code (use brutal-code-reviewer), prose (use brutal-writing-editor), resumes (use brutal-resume-reviewer), pitches (use brutal-pitch-reviewer), or designs (use brutal-design-critic) when those skills clearly apply. Do not use for genuine emotional support or when the user has not explicitly asked to be roasted.
- ▌ Bs Detector · premdevaiA rigorous BS auditor for any text, claim, pitch, marketing copy, press release, executive memo, viral tweet, news article, research summary, or "study finds" headline. Flags vague language, weasel words, unfalsifiable claims, missing evidence, suspicious metrics, and rhetoric that performs authority without earning it. Use this skill whenever the user shares text and asks "is this BS", "fact-check the framing", "audit this claim", "what's wrong with this argument", "is this real or marketing", "spot the weasel words", or wants help separating substance from performance. Trigger on phrases like "smell test this", "is this legit", "find the BS", "audit this". Different from brutal-writing-editor (which edits for craft) — this skill audits for honesty and rigor. Do not use for code (brutal-code-reviewer) or design (brutal-design-critic).
- ▌ Devils Advocate · premdevaiA rigorous adversarial thinking partner that argues the strongest possible case AGAINST any plan, decision, idea, hypothesis, business move, hire, technical choice, or strategy the user proposes. This is steelmanned opposition, not roasting — the goal is to surface the best counterargument the user will face from a smart skeptic. Use this skill whenever the user says "play devil's advocate", "argue against this", "what's the case against", "talk me out of this", "stress-test this idea", "poke holes in my plan", "convince me not to do this", or shares a decision and asks for the strongest opposing view. Trigger even when phrased casually ("am I wrong about this?", "what would a critic say?"). Do not use for general roasting (use roast-mode), code review (brutal-code-reviewer), or when the user wants encouragement rather than opposition.
- ▌ Assumption Auditor · premdevaiA structured tool for surfacing the unspoken assumptions inside any plan, argument, decision, business model, technical design, or strategy. Lists every assumption the user is making (often without realizing it), distinguishes load-bearing from decorative ones, and flags which ones are most dangerous if wrong. Use this skill whenever the user says "what am I assuming", "audit my assumptions", "what's load-bearing here", "what could be wrong about my thinking", "what am I taking for granted", or shares a plan and asks for an assumption check. Trigger on phrases like "find my blind spots", "what assumptions does this rest on". Different from devils-advocate (argues against the plan) and pre-mortem (assumes failure and finds why) — this skill enumerates the hidden premises so the user can interrogate them. Do not use for general critique (use roast-mode) or code review (brutal-code-reviewer).
- ▌ Brutal SEO Reviewer · premdevaiA merciless SEO specialist that shreds your HTML, metadata, and content structure. Roasts keyword stuffing, broken heading hierarchies, missing alts, and lazy semantic markup. Use this skill when the user asks you to review a page for SEO, check their meta tags, audit their headings, or asks why their page won't rank. Trigger this for phrases like "check my SEO", "is this page optimized", "review my meta tags", or "roast my markup". Do not use for generic UI design critiques (use brutal-design-critic) or Core Web Vitals (use brutal-web-vitals-reviewer).
- ▌ Code Quality Review · premdevaiPerform deep, adversarial code review focusing on correctness, maintainability, performance, and long-term scalability — not just style. Use this skill when the user explicitly asks for an adversarial review, mentions code quality, maintainability, scalability, or performance. Do not use for generic tutoring or polite advice.
- ▌ Brutal Code Reviewer · premdevaiA no-mercy, hyper-critical senior engineer code reviewer. Use this skill whenever the user asks for a code review, asks "what's wrong with this code", pastes code and asks for feedback, requests a critique, asks to roast/destroy/tear apart their code, asks for harsh/brutal/honest feedback, or wants engineering-grade scrutiny on a snippet, file, PR, or diff. Trigger this even when the user just shares code with phrases like "thoughts?", "review this", "be harsh", "no sugarcoating", or "tell me what sucks". Do not use for friendly tutoring, beginner hand-holding, generic explanations of how code works, README/docs reviews (use brutal-readme-reviewer), or commit message reviews (use brutal-commit-message-reviewer).
- ▌ Brutal Design Critic · premdevaiA merciless design critic for UI, UX, landing pages, marketing sites, Figma mockups, app screens, dashboards, posters, logos, and any visual interface. Use this skill whenever the user shares a design (screenshot, link, Figma, image, HTML, or description) and asks for a review, critique, audit, redesign feedback, or honest take, OR says things like "rip my landing page", "what's wrong with this UI", "is my design bad", "destroy my Figma", "be harsh on this layout". Trigger on phrases like "review my design", "critique my landing page", "is this hero section bad". Do not use for code-level review of frontend implementation (use brutal-code-reviewer), prose copy alone (use brutal-writing-editor), or pitch decks (use brutal-pitch-reviewer).
- ▌
- ▌ Brutal Pitch Reviewer · premdevaiA merciless reviewer for pitch decks, investor memos, one-pagers, startup pitches, fundraising narratives, and product launch pitches. Destroys vague TAM math, undefined moats, "AI-powered" as a feature, the four-slide deck pretending to be ten, and every other way founders avoid making real claims. Use this skill whenever the user shares a pitch deck, slide deck, investor email, fundraising one-pager, or YC-style application and asks for a review, critique, harsh feedback, or "would an investor pass on this". Trigger on phrases like "review my deck", "critique my pitch", "rip my one-pager", "is my pitch bad", "would you invest in this". Do not use for general design feedback (use brutal-design-critic), prose articles (brutal-writing-editor), or resumes (brutal-resume-reviewer).
- ▌ Brutal Writing Editor · premdevaiA merciless prose editor that destroys filler, weak verbs, throat-clearing intros, passive voice abuse, "in today's world" energy, hedging, and every other cliché writers cling to. Use this skill whenever the user shares prose — essays, blog posts, articles, newsletters, marketing copy, op-eds, fiction excerpts, LinkedIn posts, Substack drafts, book chapters — and asks for an edit, critique, review, or feedback, OR says things like "be harsh on this draft", "rip into my essay", "is this any good", "edit this", "tell me what's bad", "make this not suck". Trigger even when the request is short ("thoughts on this draft?") if prose is attached. Do not use for code (brutal-code-reviewer), pitch decks (brutal-pitch-reviewer), resumes (brutal-resume-reviewer), READMEs (brutal-readme-reviewer), or emails (brutal-email-reviewer) when those skills clearly apply.
- ▌ Brutal Readme Reviewer · premdevaiA merciless reviewer for README files, project documentation, API docs, and developer-facing technical writing. Destroys missing install steps, broken examples, "TODO" sections shipped to production, badge soup, and READMEs that explain everything except what the project actually does. Use this skill whenever the user shares a README, docs page, contributing guide, or API reference and asks for a review, critique, or harsh feedback, OR says things like "review my readme", "is my readme bad", "rip my docs", "why does no one star my repo". Trigger on phrases like "review my project docs", "critique my readme", "audit my docs". Different from brutal-code-reviewer (which reviews code) and brutal-writing-editor (which reviews prose) — this skill reviews technical documentation specifically.
- ▌ Brutal Resume Reviewer · premdevaiA merciless resume, CV, cover letter, and LinkedIn profile reviewer. Destroys "responsible for" bullets, vague metrics, buzzword soup ("passionate self-starter"), bloat, and any section that adds length without value. Use this skill whenever the user shares a resume, CV, cover letter, or LinkedIn profile and asks for a review, critique, edit, or harsh feedback, OR says things like "review my resume", "is my resume bad", "rip my CV", "destroy my LinkedIn", "why am I not getting interviews". Trigger on resume-shaped documents even with casual asks. Stay strictly on the document. Do not roast the person or their career — this skill targets the writing on the page, never the human's history or worth. Do not use for general prose (brutal-writing-editor), pitch decks (brutal-pitch-reviewer), or emails (brutal-email-reviewer).
- ▌ Interview Prep Destroyer · premdevaiA ruthless mock interviewer who tears apart practice interview answers, resumes-in-context, behavioral stories, and technical explanations. Use this skill whenever the user asks to "prep for an interview", "practice behavioral questions", "review my STAR answer", "mock interview me", "critique my interview answer", "help me prepare for a tech interview", asks "how would I answer this interview question", shares an interview question and their draft answer, or says "interview me". Trigger even when the user casually shares a question like "they asked me about a time I led a project — here's what I said." Do not use for resume document reviews (use brutal-resume-reviewer), pitch deck reviews (use brutal-pitch-reviewer), or general roasting (use roast-mode).
- ▌ Brutal Web Vitals Reviewer · premdevaiA merciless performance engineer that tears apart Core Web Vitals (LCP, INP, CLS). Demands hard data, rejects vanity metrics, and roasts lazy optimizations. Use this skill when the user asks to review web performance, LCP, INP, CLS, Lighthouse scores, or asks why their site is slow. Trigger this for phrases like "review my web vitals", "why is my LCP so bad", "fix my INP", or "roast my performance". Do not use for generic backend performance reviews (use brutal-architecture-reviewer).
- ▌ Brutal Architecture Reviewer · premdevaiA merciless senior systems architect who tears apart system designs, architecture diagrams, service boundaries, data flows, and infrastructure decisions. Use this skill whenever the user asks to "review my architecture", "critique my system design", "is this design scalable", shares a system diagram or architecture doc asking for feedback, asks about microservices vs monolith trade-offs for their specific system, or requests scrutiny on their tech stack choices, database selection, API boundaries, caching strategy, or infrastructure topology. Trigger even when the user shares a high-level design doc with "thoughts?" or "what would you change?". Do not use for code-level reviews (use brutal-code-reviewer), UI/UX design reviews (use brutal-design-critic), or API documentation reviews (use brutal-readme-reviewer).
- ▌ Production Readiness Reviewer · premdevaiA systematic, no-fluff production readiness reviewer that scans diffs and code changes across 10 critical dimensions and emits a compact 3-table verdict. Use this skill whenever the user asks for a "pre-push check", "production readiness review", "is this safe to ship", "ready to merge?", "pre-merge check", "ship check", "deploy review", or anything asking whether code is production-safe. Trigger automatically on phrases like "pre-push", "prod ready", "ready to ship", "safe to deploy", "can I merge this", "last check before PR", "is this prod-safe", or "run a ship check". Do not use for general code reviews (use brutal-code-reviewer), architecture reviews (use brutal-architecture-reviewer), or performance-only reviews (use brutal-web-vitals-reviewer).
- ▌ Brutal Commit Message Reviewer · premdevaiA merciless reviewer for git commit messages, PR titles, and PR descriptions. Destroys "fix stuff", "wip", "asdf", "final", "final final", subject lines that don't say what changed, descriptions that summarize the diff instead of explaining the why, and 50-line subject lines. Use this skill whenever the user shares commit messages, a git log, PR titles, or PR descriptions and asks for a review or critique, OR says things like "are my commit messages bad", "review my git log", "rip my PR titles", "is my commit history embarrassing". Trigger on phrases like "audit my commits", "are these commit messages bad", "review this PR description". Different from brutal-code-reviewer (reviews code) and brutal-readme-reviewer (reviews docs) — this skill is specifically for git history and PR metadata.