Coevolution extension

src/repo_registry/repo_scanning/scanner.py

@@ -69,6 +69,23 @@ MANIFEST_FRAMEWORK_HINTS = {
     },
 }
 
+LLM_PROVIDER_HINTS = {
+    "openrouter": "OpenRouter",
+    "anthropic": "Anthropic",
+    "claude": "Claude",
+    "openai": "OpenAI",
+    "gemini": "Gemini",
+    "google-generativeai": "Gemini",
+}
+
+LLM_CREDENTIAL_HINTS = {
+    "OPENROUTER_API_KEY": "OpenRouter API key",
+    "ANTHROPIC_API_KEY": "Anthropic API key",
+    "OPENAI_API_KEY": "OpenAI API key",
+    "GEMINI_API_KEY": "Gemini API key",
+    "GOOGLE_API_KEY": "Google API key",
+}
+
 
 @dataclass(frozen=True)
 class FactCandidate:
@@ -102,6 +119,7 @@ class DeterministicScanner:
         facts.extend(self._classified_file_facts(files, root))
         facts.extend(self._framework_facts(files, root))
         facts.extend(self._interface_facts(files, root))
+        facts.extend(self._llm_provider_facts(files, root))
 
         return ScanResult(
             source_path=str(root),
@@ -223,6 +241,96 @@ class DeterministicScanner:
                 facts.append(FactCandidate("interface", "possible API surface", relative))
         return facts
 
+    def _llm_provider_facts(self, files: list[Path], root: Path) -> list[FactCandidate]:
+        facts: list[FactCandidate] = []
+        seen: set[tuple[str, str, str]] = set()
+        for path in files:
+            if path.suffix.lower() not in {
+                ".py",
+                ".ts",
+                ".js",
+                ".json",
+                ".toml",
+                ".yaml",
+                ".yml",
+                ".md",
+                ".txt",
+                ".env",
+            } and not path.name.lower().startswith(".env"):
+                continue
+            try:
+                text = path.read_text(encoding="utf-8", errors="ignore")
+            except OSError:
+                continue
+            lower_text = text.lower()
+            relative = path.relative_to(root).as_posix()
+            for needle, provider in LLM_PROVIDER_HINTS.items():
+                if needle not in lower_text:
+                    continue
+                self._append_once(
+                    facts,
+                    seen,
+                    FactCandidate(
+                        kind="llm_provider",
+                        name=provider,
+                        path=relative,
+                        value=needle,
+                        metadata={"source": "provider_hint"},
+                    ),
+                )
+            for env_name, label in LLM_CREDENTIAL_HINTS.items():
+                if env_name.lower() not in lower_text:
+                    continue
+                self._append_once(
+                    facts,
+                    seen,
+                    FactCandidate(
+                        kind="credential_config",
+                        name=label,
+                        path=relative,
+                        value=env_name,
+                        metadata={"source": "environment_variable"},
+                    ),
+                )
+            if any(term in lower_text for term in ("provider_registry", "providers =", "adapter")):
+                if any(needle in lower_text for needle in LLM_PROVIDER_HINTS):
+                    self._append_once(
+                        facts,
+                        seen,
+                        FactCandidate(
+                            kind="provider_registry",
+                            name="LLM provider registry",
+                            path=relative,
+                            metadata={"source": "provider_registry_hint"},
+                        ),
+                    )
+            if "fallback" in lower_text and any(
+                needle in lower_text for needle in LLM_PROVIDER_HINTS
+            ):
+                self._append_once(
+                    facts,
+                    seen,
+                    FactCandidate(
+                        kind="fallback_policy",
+                        name="LLM provider fallback policy",
+                        path=relative,
+                        metadata={"source": "fallback_hint"},
+                    ),
+                )
+        return facts
+
+    def _append_once(
+        self,
+        facts: list[FactCandidate],
+        seen: set[tuple[str, str, str]],
+        fact: FactCandidate,
+    ) -> None:
+        key = (fact.kind, fact.name, fact.path)
+        if key in seen:
+            return
+        seen.add(key)
+        facts.append(fact)
+
     def _python_interface_facts(self, path: Path, relative: str) -> list[FactCandidate]:
         facts: list[FactCandidate] = []
         try: