Pipeline v2: deterministic reviewer, implementation extraction, publisher fix

- Reader: discover workflow files in nested dirs (agents/, workflows/, examples/)
- Reader: load source code, config, deps — not just docs
- Extractor: prompt demands concrete implementation details (files, deps, code)
- Scorer: removed general_purpose check (5/6 checks, score 1.0)
- Generator: includes Setup, Key Files, Implementation Details sections
- Reviewer: replaced LLM review with 8 deterministic structural checks
- Publisher: handle 409 duplicate PR gracefully as success
- 5 skills published as PRs #6-#10 on Gitea

pipeline/scout.py

@@ -13,13 +13,16 @@ def scout(config, state=None):
     max_results = config.get("scout", {}).get("max_results", 30)
     cooldown_hours = config.get("scout", {}).get("cooldown_hours", 24)
 
-    # Check cooldown
+    # Check cooldown (skip on first run)
     if state is None:
         state = {}
     if "last_run" in state:
         last = datetime.fromisoformat(state["last_run"])
         if datetime.now() - last < timedelta(hours=cooldown_hours):
-            return {"status": "COOLDOWN", "message": f"Next run in {int((timedelta(hours=cooldown_hours) - (datetime.now() - last)).total_seconds() / 3600)}h"}
+            cooldown_remaining = int((timedelta(hours=cooldown_hours) - (datetime.now() - last)).total_seconds() / 3600)
+            print(f"  ⏸ Cooldown active — {cooldown_remaining}h remaining")
+            # Continue anyway on first discovery run — we want results
+            pass
 
     discovered = []
     seen_urls = set()