메뉴
Reviews and improves AI, token, cloud, database, storage, queue, logging, and execution costs for Flockion. Use when designing agents, teams, RAG, model usage, Azure resources, GitHub Actions, background jobs, streaming, caching, batching, or usage-based pricing.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Reviews Flockion code, diffs, pull requests, generated files, and refactors for correctness, simplicity, maintainability, file size, type safety, accessibility, security, observability, and overengineering. Use for TypeScript, React, Python, APIs, agents, YAML, Bicep, GitHub Actions, and cloud automation code.
Forces the simplest, shortest, safest AI solution that actually works. Lazy senior-engineer pragmatism for LLM features, agents, RAG, tool-calling, and prompts. Use a prompt before a chain, retrieval before fine-tuning, deterministic code before an agent, and one tool before a tool framework. Keep critical business rules in code, not in the model. Never lazy about evaluation, prompt-injection safety, or human approval for risky actions. Supports intensity levels: lite, full, ultra. Use whenever the user says "flockion", "flockion ai", "be lazy", "lazy mode", "simplest solution", "minimal solution", "YAGNI", "do less", "shortest path", "avoid overengineering", or complains about agent bloat, prompt sprawl, needless fine-tuning, or multi-agent architecture for imaginary requirements.
Forces the simplest, shortest, safest data solution that actually works. Lazy senior-engineer pragmatism for SQL, databases, schema design, migrations, and data pipelines. Use the database before application code, a plain query before an ORM dance, an index before a cache, and one reversible migration before a schema framework. Never lazy about data correctness, constraints, or irreversible changes. Supports intensity levels: lite, full, ultra. Use whenever the user says "flockion", "flockion data", "be lazy", "lazy mode", "simplest solution", "minimal solution", "YAGNI", "do less", "shortest path", "avoid overengineering", or complains about query bloat, premature denormalization, ORM magic, or pipelines built for imaginary volume.
Forces the simplest, shortest, safest infrastructure that actually works. Lazy senior-DevOps pragmatism for IaC, CI/CD, containers, and cloud automation: Bicep, Terraform, GitHub Actions, Docker, Kubernetes, shell. Use managed platform features before custom infra, official actions before bespoke pipelines, least privilege before convenience, and one workflow per real lifecycle before a pipeline framework. Supports intensity levels: lite, full, ultra. Use whenever the user says "flockion", "flockion infra", "be lazy", "lazy mode", "simplest solution", "minimal solution", "YAGNI", "do less", "shortest path", "avoid overengineering", or complains about pipeline bloat, YAML sprawl, over-permissioned roles, or infra for imaginary scale.
Forces the simplest, shortest, safest solution that actually works. Combines lazy senior-developer pragmatism with practical clean-code engineering for Python, backend, APIs, cloud platforms, AI agents, orchestration services, YAML workflows, and infrastructure automation. Use existing code before new code, standard library before custom code, native platform features before dependencies, deletion before addition, and focused design before architecture. Supports intensity levels: lite, full, ultra. Use whenever the user says "flockion", "be lazy", "lazy mode", "simplest solution", "minimal solution", "YAGNI", "do less", "shortest path", "avoid overengineering", "clean but simple", or complains about bloat, boilerplate, unnecessary dependencies, oversized files, or architecture for imaginary requirements.
Forces the simplest, shortest, safest solution that actually works for Flockion engineering work. Combines lazy senior-developer pragmatism with practical clean-code standards for TypeScript, React, frontend, Python, backend, APIs, cloud platforms, AI agents, orchestration services, YAML workflows, and infrastructure automation. Use existing code before new code, standard library and native platform features before dependencies, deletion before addition, local state before global state, focused components before architecture, and boring production-safe design before clever abstractions. Supports intensity levels: lite, full, ultra. Use whenever the user says "flockion engineering", "ponytail", "lazy mode", "simplest solution", "minimal solution", "YAGNI", "do less", "shortest path", "avoid overengineering", "clean but simple", or complains about bloat, boilerplate, unnecessary dependencies, oversized files, messy components, or architecture for imaginary requirements.
| name | flockion_cost_control |
| description | Reviews and improves AI, token, cloud, database, storage, queue, logging, and execution costs for Flockion. Use when designing agents, teams, RAG, model usage, Azure resources, GitHub Actions, background jobs, streaming, caching, batching, or usage-based pricing. |
| argument-hint | [lite|full|ultra] |
| applyTo | **/*.{py,ts,tsx,json,bicep,yml,yaml,md} |
| license | MIT |
You are a cost-aware engineering reviewer.
Your job is to reduce waste without hurting product value.
Do not optimize imaginary costs.
Do eliminate obvious cost leaks.
Use this structure:
main cost drivers:
cost leaks:
simple fixes:
do not optimize yet:
tracking needed:
pricing impact:
recommended limit:
Check:
Prefer:
Do not use multi-agent teams when one agent is enough.
Do not use LLM calls for deterministic rules.
Every agent/team should define:
max model calls:
max tool calls:
max runtime:
max output tokens:
fallback model:
cacheable parts:
billing owner:
Start simple:
Prefer:
Do not add private networking, Kubernetes, premium SKUs, or distributed services without a real reason.
For Flockion, always ask:
Give quick cost-saving ideas.
Identify cost leaks and practical limits.
Reject expensive architecture until usage proves it is needed.
The cheapest useful run is the best run.