init: llm-wiki-haness 하네스 설계

harness/source/agents/wiki-adversarial-reviewer.json

@@ -0,0 +1,30 @@
+{
+  "schema_version": 1,
+  "source_kind": "agent",
+  "id": "wiki-adversarial-reviewer",
+  "description": "Use AFTER a wiki research/audit draft (master report + per-file findings, typically from wiki-research-lane output) exists, and BEFORE the final priority recommendations are locked in. Takes the draft and attempts to FALSIFY each finding via Practicality / Overclaim / Assumption checks. Recommends KEEP / DOWNGRADE / REJECT per finding. Read-only. Use when the draft has ≥5 findings — its purpose is to break the rubber-stamp loop that occurs when the same agent self-reviews.",
+  "heading_locale": "mixed",
+  "body_path": "harness/source/agents/bodies/wiki-adversarial-reviewer.md",
+  "targets": [
+    {
+      "kind": "plugin-agent-md",
+      "path": ".agents/plugins/wiki-superpowers/agents/wiki-adversarial-reviewer.md"
+    },
+    {
+      "kind": "claude-agent-md",
+      "path": ".claude/agents/wiki-adversarial-reviewer.md"
+    },
+    {
+      "kind": "codex-agent-md",
+      "path": ".codex/agents/wiki-adversarial-reviewer.md"
+    },
+    {
+      "kind": "codex-agent-toml",
+      "path": ".codex/agents/wiki-adversarial-reviewer.toml"
+    },
+    {
+      "kind": "antigravity-agent-json",
+      "path": ".agents/agents/wiki-adversarial-reviewer/agent.json"
+    }
+  ]
+}