feat(skills): give some refresh to our skills

[OisinKyne]

Summary

→ BUILD → EVALUATE → PRICE → SELL → MEASURE → ITERATE/KILL) with exact commands wiring together discovery (scout competitors), buy-x402 (probe their prices), agent-factory (build the child), sell (offer status), and ethereum-networks (read revenue on-chain). Its centrepiece is the evaluate gate: 10 hardest buyer questions vs. a raw-model baseline, list only on ≥7/10 wins — the honest answer to "why wouldn't the buyer just ask GPT?". Includes value-anchored pricing ladders, operator handoffs (tunnel/branding/registration), and a kill-fast rule.

• New skill swap — quote-first Uniswap V3 recipes (QuoterV2 → approve → exactInputSingle → unwrap) on Base and mainnet via the remote-signer, with slippage floors, mandatory user confirmation, and OBOL two-hop guidance. Every address was verified on-chain via eth_getCode today — I didn't trust my memory or even the existing docs.
• addresses skill: new references/obol-payments.md (USDC + OBOL per supported chain from tokens.go/chains.go, Permit2, ERC-8004 registries, all verified today) listed first in the index.
• Trigger sharpening on agent-factory, sell, monetize-guide — explicit "use when" phrasing plus cross-links so a weak model routes between the plumbing skill, the walkthrough, and the business layer correctly.
• README skill tables rewritten to the real 23 skills; CLAUDE.md count fixed; new skills added to embed_skills_test.go coreSkills (tests pass).

[bussyjd]

Closing as superseded by integration/v0.13.0-rc0 @ a29c826 (rolled up in #716 / release v0.13.0-rc0). All commits of this PR are contained in the RC tip (verified by ancestry/patch-id audit).