feat: add dbs-offer skill for offer packaging and conversion path diagnosis

[Sunnyeung369]

Summary

This PR adds /dbs-offer, a new skill for diagnosing whether a fuzzy idea, service, AI project, or content direction can become a sellable, deliverable, and verifiable offer.

Why

Current dbskill already covers diagnosis, benchmark, content, hook, goal, decision, learning, state management, and agent migration.

One common gap appears between business diagnosis and content execution:

Users may know the direction, but still cannot clearly answer:

• Who pays?
• Why now?
• How much?
• What gets delivered?
• How is it verified?
• What is the next conversion action?

/dbs-offer is designed to fill this gap.

What changed

• Added skills/dbs-offer/SKILL.md - Complete offer diagnosis skill with 6 core philosophies, 6-phase diagnosis process, 6-dimensional scoring, and AI project specific rules
• Updated /dbs router to include offer routing in the skill selection menu
• Updated README tool list to include /dbs-offer
• Updated README workflow to show benchmark → offer → content flow
• Added 知识库/Skill知识包/offer_产品化与成交路径.md - Standalone knowledge base with offer methodology, pseudo-offers, real offer characteristics, product tiers, and diagnostic templates

Flow integration

diagnosis → offer
benchmark → offer
content → offer
offer → content
offer → decision
offer → save

Notes

This PR does not change VERSION or release files.

No external dependencies added.

No private commercial assets included - All content is generic frameworks and methodologies applicable to any business or project.

No project refactoring - This is a pure addition that integrates with existing skill ecosystem without modifying core architecture.

Follows existing dbskill writing style - Short, direct, diagnostic-focused, anti-self-hype.

Skill routing updates

The /dbs main router now includes:

• 想设计产品、定价、成交路径、说"我卖什么" → /dbs-offer

Knowledge base

Added offer_产品化与成交路径.md with:

1. Offer definition (payer + pain + deliverable + price + next action)
2. Common pseudo-offers to avoid
3. Real offer characteristics (5 features)
4. Four-tier product gradient (free entry → low validation → medium product → high ticket)
5. AI project specific rules (6 rules)
6. Content to conversion path
7. Offer diagnostic templates

Testing

The skill has been structured to integrate with existing skill recommendations:

• diagnosis finds direction成立 but product unclear → recommends offer
• content finds no product or CTA can't capture conversion → recommends offer
• benchmark finds benchmark but doesn't know how to package their service → recommends offer
• offer established and needs content acquisition → recommends content
• offer involves major pricing or transformation choice → recommends decision
• offer diagnosis complete → recommends save