Startup Competitors
Vícefázový výzkum konkurence do battle cards, cenového prostředí a matice funkcí
Otestováno · Funguje
Co umí
Dovednost pro konkurenční zpravodajství, která profiluje 5-8 competitors across three research waves (product/pricing, customer sentiment, GTM/growth signals), then synthesizes battle cards, a pricing landscape, a feature matrix, and a strategic report with confidence ratings and declared data gaps. Spouští se on requests like "who are my competitors", "competitive analysis", "battle cards", or "competitor pricing comparison". Uses live web search when available and falls back to a labeled knowledge-based mode when it is not.
Testovací report
Načteno SKILL.md a spot-checked 3 referenced files (research-scaling/synthesis/honesty-protocol) — all HTTP 200; frontmatter has name+description, no hardcoded paths or security smells. Trigger phrasings — SHOULD (all judged load): "Who are my competitors in the AI note-taking space?", "Build me battle cards for my top 3 competitors", "Do a pricing comparison across the PM tools we're up against"; should NOT (correctly declined): "Design my startup's brand identity and logo", "Write a cold email to a prospect at Acme" — 5/5. Output test: produced a Calendly battle card with no skill (sc_baseline.md) vs strictly following the battle-card template + honesty protocol in Knowledge-Based mode (sc_skill.md). Concrete difference: baseline asserts "generous free tier" and invents no review evidence with zero sourcing or numbers; the skill version adds [Data]/[Estimate]/[Assumption]/[Opinion] labels, quantifies funding/pricing with >12mo staleness flags, uses objection->response format, and writes "DATA GAP: No verified verbatim review quote available" instead of fabricating a customer quote, plus Red/Yellow Flags and Sources. I could not exercise the skill's core live-web multi-agent research engine, only its synthesis/honesty layer, so output is a real but bounded improvement (7).
Testováno: 2026-07-21 · Claude Code 2.x (agent harness)
Instalace
git clone --depth 1 https://github.com/ferdinandobons/startup-skill.git /tmp/startup-competitors-src mkdir -p ~/.claude/skills cp -R /tmp/startup-competitors-src/startup-competitors ~/.claude/skills/startup-competitors # SKILL.md lands at ~/.claude/skills/startup-competitors/SKILL.md # Needs WebSearch for full value; without it, runs in labeled Knowledge-Based fallback mode. # In Claude Code it spawns parallel research agents; on Claude.ai it runs the same waves sequentially. # Plugin-marketplace alternative: repo ships .claude-plugin/marketplace.json (plugin "startup", 4 skills).
Příkazy a ukázkové prompty
/startup-competitorsVícefázový výzkum konkurence do battle cards, cenového prostředí a matice funkcí
Skilly se spouštějí běžnými požadavky — žádné příkazy k zapamatování. Po instalaci ho aktivují prompty jako tyto (anglicky):
Analyze my competitors' pricing and positioningBuild battle cards for our top three competitorsResearch who we're competing against in this market