Design tool systems for AI agents — core/extended/deferred tiers, eval compute, batch tools, trust metadata. Use when building agent tools, reviewing tool architecture, or reducing tool count.
Use when "find first customers", "первые клиенты", "early adopters", "design partners", "who would buy this", "prospect list", or need an evidence-backed shortlist of potential first customers from public signals. Do NOT use for community thread replies (/community-outreach), market research (/research), or GTM strategy (/launch).
Use when "scaffold project", "create new project", "start new app", "bootstrap project", "set up from PRD", or need project from PRD + stack template. Do NOT use for planning features (/plan) or PRD generation (/validate).
Use when needing methodology reference, asking about principles, harness engineering, SGR, launch playbook, agent memory, decision frameworks, or any topic covered in the solopreneur knowledge base. Also use when the user asks "what do we know about X" or "how does our methodology handle X".
Use when "design schemas", "structured output", "agent loop", "SGR", "constrained decoding", "tool dispatch", "Pydantic schema for LLM", or need to design a schema-guided reasoning pipeline for an agent or API. Do NOT use for general code review (/review) or planning (/plan).
Use when "build it", "start building", "execute plan", "implement tasks", "ship it", track ID referenced, or plan tasks need execution. Do NOT use for planning (/plan) or scaffolding (/scaffold).
Use when "deploy it", "push to production", "set up hosting", or after /build completes and project needs hosting. Do NOT use before build is complete.
Use when "audit KB", "check frontmatter", "find broken links", "tag cleanup", "knowledge base quality", or docs need health check. Do NOT use for SEO audits (/seo-audit) or code reviews (/review).