Add skill for cluster and replication

[valkirilov]

Why

Continues the agentskills.io migration of redis-development's rules into focused spec-compliant skills (see #20–#25).

This PR carves out the clustering and replication topic into its own skill so an agent has sharp triggers for "Redis Cluster", "hash tag", "CROSSSLOT", "read replica" instead of fishing through a 33-rule grab-bag.

Spec reference

• Spec: https://agentskills.io/specification
• Layout: SKILL.md + references/.

What's in this PR

A single commit, additive only — no files deleted from skills/redis-development/rules/. The legacy redis-development skill continues to include the two cluster-* rules unchanged, so downstream plugin consumers keep working through the migration. Duplicates are removed in the final cleanup PR.

Two rules folded into `redis-clustering`:

SKILL.md has two sections, framed around the two failure modes that bite most new cluster users:

• Hash tags for multi-key operations — only the {...} portion hashes to a slot; MGET / SDIFF / pipelines / transactions need all keys on the same slot or they error out with CROSSSLOT. Includes the guidance to scope tags to the meaningful entity ({user:1001}) rather than bare identifiers ({1001}) that collide across namespaces.
• Read replicas for read-heavy workloads — read_from_replicas=True for Redis Cluster and standalone replica setups, with the eventually-consistent caveat made concrete: fine for caches / analytics / dashboards, not for balances or idempotency state.

Validation

• skill-validator check skills/redis-clustering → passed, 0 warnings.
• npm run validate → rules, claude-plugins, cursor-plugins all green.
• Branched from main, independent of the other migration PRs. Can merge in parallel.

Out of scope (follow-ups)

• Removing the duplicate cluster-*.md source rules — happens in the final cleanup PR.
• Authoring evals for clustering topics — separate work after the structural migration completes.

🤖 Generated with Claude Code

---

Note

Low Risk
Low risk because this PR is additive documentation/content only and does not change runtime code paths or behavior.

Overview
Introduces a new spec-compliant redis-clustering skill (SKILL.md + references/) that consolidates Redis Cluster key-design guidance around hash tags for multi-key operations (avoiding CROSSSLOT errors in pipelines/transactions/scripts) and routing reads to replicas for read-heavy workloads, including the eventual-consistency caveats.

Adds two standalone reference docs (references/hash-tags.md, references/read-replicas.md) with concrete Python/Java examples and links to official Redis docs.

^{Reviewed by Cursor Bugbot for commit d6e6f4b. Bugbot is set up for automated code reviews on this repo. Configure here.}

[jit-ci]

🛡️ Jit Security Scan Results

✅ No security findings were detected in this PR

---

^{Security scan by Jit}