docs(consent): rewrite with risk profiles, benefits, and command descriptions

[TeoSlayer]

Summary

Complete rewrite of /docs/consent to lead with what actually matters to users making an opt-out decision:

• Risk level badge (Low / Medium) prominently displayed for each feature
• Benefits breakdown per stakeholder — you / the network / developers — so users understand why each feature is on by default before deciding whether to opt out
• Explicit risk profile per feature: what data is actually collected/received, the realistic threat model, and what you're accepting in plain English
• "Who should turn this off" concrete guidance rather than just "you can opt out"
• Full command reference with inline comments explaining exactly what each command does — not just the syntax
• Skill injection three-mode table with a "Best for" column so users can pick the right risk/convenience trade-off
• Sandbox mode documented as security hardening, not a privacy feature
• Risk-level styles (risk-low-badge, risk-medium-badge) scoped to the page so they don't bleed globally

Test plan

• Page renders at /docs/consent with correct sidebar active state
• Risk badges display with correct color (green=low, amber=medium)
• All 7 anchor links in the TOC scroll to the right section
• Commands in code blocks are accurate to the current binary

🤖 Generated with Claude Code