dbs-benchmark: Stop Self-Talk, Start Copying

Where this fits

You've got a new SaaS idea, but you've spent three hours debating whether it matches your skills, your passions, or your existing network. Or you're running a small e-commerce store and want to know which competitor to actually emulate, not just admire. The dbs-benchmark skill cuts through that self-referential fog. It forces you (or your agent) to answer one question: Can this business be executed? If yes, you copy it—exactly. This skill is for developers building agents that advise on business strategy or product direction, where the hardest bottleneck isn't data—it's the user's ego.

Why agents benefit

• Enforces a ruthless filter chain: The five screens (profit, 10x income, exact copy granularity, etc.) are deterministic. An agent can step through them programmatically, rejecting candidates that fail, without negotiating with the user's emotional attachments.
• Interrupts unproductive self-talk: The skill instructs the agent to literally stop the user when they say "I want something that fits me." That's a natural dialogue-handling rule, not a vague nicety.
• Computes profit signals from prices and volumes: Instead of asking the user for vague “high-margin” vibes, the agent estimates revenue from product pricing and mock sales volumes, making the comparison concrete.
• Generates a shortlist of copyable targets: After filtering, the agent outputs a ranked list of businesses to imitate, complete with specific granular details to match—right down to sock threads, if that's the level of fidelity required.
• Works with zero domain configuration: Trigger it with /dbs-benchmark or a natural phrase. No API keys, no training—just the skill's instructions and your agent's existing reasoning.

Practical scenarios

User with an existing online course business

The agent asks what the user is doing. The user says they sell a $49 course on gardening. The agent scans the landscape for similar courses that price at least 10x higher (say $490+), verifies they're still selling via mock checkout flows, and presents a concrete competitor to dissect—including their sales funnel, testimonials, and even reflection on content length. The user stops rationalizing why they "don't have time" and starts cloning the winning structure.

User without a business but "full of ideas"

They say they want to "do something meaningful." The agent interrupts: "We don't discuss you. We discuss the business." Then it pulls from a dataset of low-effort, high-margin niches (e.g., print-on-demand accessories, micro-SaaS admin tools) and runs the five-filter on each. It returns one clear winner with a detailed copying plan. The user is shocked but ultimately grateful.

Team evaluating a pivot

A startup team asks "who should we copy now?" The agent filters out their current B2B SaaS (low profit per user) and suggests a B2C high-repeat model with proven peers. The team realises the pivot is not about "building something new" but about "copying something that works better."

Add it to your agent workflow

Install with npx, then trigger it inside any agent conversation:

npx skills add Pluviobyte/rnskill --skill dbs-benchmark

Then simply say: 帮我找对标 or who should I copy?. The agent will start with the Phase 1 question about your current state and guide you through the filtering process. You can also call it programmatically in your agent loop by injecting the skill's instructions as context whenever the user's intent matches.

Compare before adopting

When evaluating this skill, compare it against other benchmarking tools you might use:

• Quality signals: Does the skill have clear validation steps (like the 10x profit rule)? Is the output actionable rather than just a list of competitors?
• Maintenance freshness: The GitHub repo has 1.3k stars, but check last commit date and open issues for stale assumptions about pricing or platforms.
• Alternatives: Generic SWOT analysis or competitor scraping tools offer breadth but lack the brutal filtering logic and the explicit “copy exact details” mandate.
• Workflow fit: Does it integrate with your agent's turn-taking style? The skill requires interruptive dialogue; make sure your agent can handle that gracefully.

Why it is worth tracking

This skill is a sharp, opinionated tool that challenges the default “do it my way” bias in early-stage decision-making. Its popularity (1.3k stars) indicates strong community alignment with the philosophy that copying is a discipline, not a shortcut. Monitor its GitHub for updates—if the author adds data sources for profit estimation or expands the filters to cover growth metrics, it becomes even more valuable. When your agent needs to make a strategic recommendation without falling into user self-reference loops, evaluate dbs-benchmark now. For similar agent skills and comparisons, see OpenAgentSkill's skill page. Tracking it today means having a tested pattern for when your users ask, "What should I actually do?"