docs(claude): add rule to fix warnings on discovery (#597)

* docs(claude): add rule to fix warnings on discovery

* fix(docs): satisfy oxfmt on CLAUDE.md

* docs(claude): drop contradictory lib/http-request rule

The SHARED STANDARDS section referenced httpJson/httpText/httpRequest from
@socketsecurity/lib/http-request, but this SDK's canonical rule (in the
SDK-SPECIFIC section) directs contributors to createGetRequest and
createRequestWithJson from src/http-client.ts. Keep the single
authoritative rule to avoid conflicting guidance.

* docs(claude): fix stale file paths and tighten warnings rule

- scripts/test.mjs → scripts/test.mts (renamed in 0e600f7)
- .config/esbuild.config.mjs → .config/esbuild.config.mts
- Clarify Judgment Protocol warnings rule: fix in-edit only, flag
  unrelated warnings instead of drive-by fixing.

Author: jdalton

.claude/skills/security-scan/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: security-scan
-description: Runs a multi-tool security scan — AgentShield for Claude config, zizmor for GitHub Actions, and optionally Socket CLI for dependency scanning. Produces an A-F graded security report.
+description: Runs a multi-tool security scan — AgentShield for Claude config, zizmor for GitHub Actions, and optionally Socket CLI for dependency scanning. Produces an A-F graded security report. Use after modifying `.claude/` config, hooks, agents, or GitHub Actions workflows, and before releases.
 user-invocable: true
 ---