feat(skills): add react useEffect ban skill

[altaywtf]

Summary

Add a shared react-ban-use-effect skill for replacing direct React useEffect usage with stronger React, server-state, and framework patterns.

Changed

• Added skills/react-ban-use-effect with SKILL, OpenAI metadata, tile metadata, replacement guidance, alternatives map, and upstream attribution.
• Added the new skill to scripts/sync/skills.json.
• Aligned autoreview upstream attribution language with the same references/upstream.md convention.

Risks

• Skill activation could overlap with ordinary React work; the description now narrows activation to direct useEffect imports/calls/replacements or no-direct-useEffect policy work.
• Replacement guidance could over-prescribe React/TanStack patterns; references include caveats for Suspense query options, React form hooks, optimistic updates, dependency-aware external sync, and cleanup.

Verification

• npx tessl@0.80.0 skill review --json skills/react-ban-use-effect -> score 100, passed, 0 warnings, 0 errors
• npx tessl@0.80.0 skill review --json skills/autoreview -> score 100, passed, 0 warnings, 0 errors
• npx tessl@0.80.0 tile lint skills/react-ban-use-effect -> valid tile; existing-style warning for agents/openai.yaml placement
• git diff --check --cached
• jq . scripts/sync/skills.json
• jq . skills/react-ban-use-effect/tile.json

Complexity

Increased slightly: this adds a new shared skill and two focused references, but keeps the entrypoint lean and one-hop discoverable.

---

Summary by cubic

Adds a shared react-ban-use-effect skill that bans direct react useEffect and guides replacements with server-state, event, and external-sync patterns. Registers the skill and updates autoreview upstream attribution.

• New Features

  • Added skills/react-ban-use-effect with SKILL.md, agents/openai.yaml, and tile.json.
  • Included replacement guide and alternatives docs with clear ladders, escape hatches, and ESLint/review enforcement notes.
  • Registered the new skill in scripts/sync/skills.json.

• Refactors

  • Aligned skills/autoreview upstream attribution and wording to use references/upstream.md.

^{Written for commit 2a684f6. Summary will update on new commits.}

Review in cubic

[cubic-dev-ai]

No issues found across 9 files

_{Re-trigger cubic}