diff --git a/skills/codealive-context-engine/SKILL.md b/skills/codealive-context-engine/SKILL.md
index 0f3c9d7..a9e4057 100644
--- a/skills/codealive-context-engine/SKILL.md
+++ b/skills/codealive-context-engine/SKILL.md
@@ -218,7 +218,7 @@ or your local file-read tool before drawing conclusions about behavior.
 Retrieves the full source code content for artifacts found via search. Use this for external repositories you cannot access locally.
 
 ```bash
-python scripts/fetch.py <identifier1> [identifier2...]
+python scripts/fetch.py <identifier1> [identifier2...] [--data-source NAME_OR_ID]
 ```
 
 | Constraint | Value |
@@ -226,12 +226,27 @@ python scripts/fetch.py <identifier1> [identifier2...]
 | Max identifiers per request | 20 |
 | Identifiers source | `identifier` field from search results |
 | Identifier format | `{owner/repo}::{path}::{symbol}` (symbols), `{owner/repo}::{path}` (files) |
+| `--data-source NAME_OR_ID` | Optional. Data source Name or Id (from a result's `Source:` line) to disambiguate an identifier indexed in more than one data source |
 
 For function-like artifacts the response includes a small **relationships
 preview** (up to 3 outgoing/incoming calls per direction). To see the full
 call graph, inheritance, or references, run `relationships.py` with the
 artifact's identifier.
 
+**Disambiguating an identifier that lives in more than one data source.** Artifact
+identifiers are unique only per data source, so the same identifier can belong to
+more than one data source. If you fetch such an identifier without `--data-source`,
+the backend returns a **409** listing the candidate data sources instead of picking
+one for you. Every listed candidate **will** resolve, so the workflow is: call without
+`--data-source` → read the 409 candidates → try one → if that data source isn't the one
+you want, try the next. To resolve it: take the
+`Source:` name or id shown next to the search result you want and pass it back —
+`python scripts/fetch.py <identifier> --data-source "backend"` (or the id).
+The same `--data-source` flag works on `relationships.py`. If a `--data-source`-scoped
+call finds nothing (the script prints a "nothing was found in data source …" hint),
+the identifier belongs to a different data source or the selector is wrong: retry with
+a different `Source:` value, or drop `--data-source` to get the 409 candidate list.
+
 ### `relationships.py` — Drill into an Artifact's Relationship Graph
 
 Returns the full call graph (incoming/outgoing calls), inheritance hierarchy
@@ -241,7 +256,7 @@ identifier and want to understand how the artifact relates to the rest of the
 codebase.
 
 ```bash
-python scripts/relationships.py <identifier> [--profile PROFILE] [--max-count N]
+python scripts/relationships.py <identifier> [--profile PROFILE] [--max-count N] [--data-source NAME_OR_ID]
 ```
 
 | Option | Description |
@@ -251,6 +266,7 @@ python scripts/relationships.py <identifier> [--profile PROFILE] [--max-count N]
 | `--profile allRelevant` | Calls + inheritance (4 groups) |
 | `--profile referencesOnly` | Symbol references |
 | `--max-count N` | Max related artifacts per relationship type (1–1000, default 50) |
+| `--data-source NAME_OR_ID` | Optional. Data source Name or Id to disambiguate an identifier indexed in more than one data source (same 409 contract as `fetch.py`) |
 | `--json` | Emit the raw JSON response instead of the formatted view |
 
 **When this adds value vs the fetch preview:**
diff --git a/skills/codealive-context-engine/scripts/fetch.py b/skills/codealive-context-engine/scripts/fetch.py
index 6986178..ae52d18 100644
--- a/skills/codealive-context-engine/scripts/fetch.py
+++ b/skills/codealive-context-engine/scripts/fetch.py
@@ -3,7 +3,7 @@
 CodeAlive Fetch - Retrieve full content for code artifacts
 
 Usage:
-    python fetch.py <identifier1> [identifier2...]
+    python fetch.py <identifier1> [identifier2...] [--data-source NAME_OR_ID]
 
 Examples:
     # Fetch a single artifact (symbol)
@@ -15,10 +15,18 @@
     # Fetch multiple artifacts
     python fetch.py "my-org/backend::src/auth.py::login" "my-org/backend::src/utils.py::helper"
 
+    # Disambiguate an identifier that exists in more than one data source
+    # (use the dataSource name or id from a search result)
+    python fetch.py "my-org/backend::src/auth.py::login" --data-source "backend"
+
 Identifiers come from semantic/grep search results (the `identifier` field).
 The format is: {owner/repo}::{path}::{symbol} (for symbols/chunks)
                {owner/repo}::{path} (for files)
 
+Pass --data-source (a data source Name or Id from a search result's `dataSource`)
+to disambiguate an identifier that exists in more than one data source. Without it,
+an ambiguous identifier returns a 409 listing the candidate data sources.
+
 Maximum 20 identifiers per request.
 """
 
@@ -83,11 +91,23 @@ def _format_relationships_preview(relationships: dict) -> list:
     return lines
 
 
-def format_artifacts(data: dict) -> str:
+def _data_source_miss_hint(data_source: str) -> str:
+    """Recovery hint when a data-source-scoped fetch returns nothing."""
+    return (
+        f'\n💡 Hint: nothing was found in data source "{data_source}". The identifier may belong to a '
+        "different data source, or the --data-source value may be wrong. Try: re-run with --data-source "
+        "set to a different candidate (use the Source name or id from your search results, or run "
+        "datasources.py), or drop --data-source entirely — an ambiguous identifier then returns a 409 "
+        "listing the candidate data sources to choose from."
+    )
+
+
+def format_artifacts(data: dict, data_source: str = None) -> str:
     """Format fetched artifacts for display."""
     artifacts = data.get("artifacts", [])
     if not artifacts:
-        return "No artifacts returned."
+        msg = "No artifacts returned."
+        return msg + _data_source_miss_hint(data_source) if data_source else msg
 
     output = []
     count = 0
@@ -119,7 +139,8 @@ def format_artifacts(data: dict) -> str:
                     has_any_relationships = True
 
     if not output:
-        return "No artifacts found."
+        msg = "No artifacts found."
+        return msg + _data_source_miss_hint(data_source) if data_source else msg
 
     output.append(f"\n({count} artifact(s))")
 
@@ -144,7 +165,26 @@ def main():
             sys.exit(1)
         sys.exit(0)
 
-    identifiers = sys.argv[1:]
+    identifiers = []
+    data_source = None
+    i = 1
+    while i < len(sys.argv):
+        arg = sys.argv[i]
+        if arg == "--data-source":
+            # Match the flag first, then require a value — otherwise a trailing "--data-source"
+            # with no value would be silently appended as an identifier.
+            if i + 1 >= len(sys.argv):
+                print("Error: --data-source requires a value.", file=sys.stderr)
+                sys.exit(1)
+            data_source = sys.argv[i + 1]
+            i += 2
+        else:
+            identifiers.append(arg)
+            i += 1
+
+    if not identifiers:
+        print("Error: At least one identifier is required.", file=sys.stderr)
+        sys.exit(1)
 
     if len(identifiers) > 20:
         print("Error: Maximum 20 identifiers per request.", file=sys.stderr)
@@ -154,11 +194,13 @@ def main():
         client = CodeAliveClient()
 
         print(f"📥 Fetching {len(identifiers)} artifact(s)", file=sys.stderr)
+        if data_source:
+            print(f"   data source: {data_source}", file=sys.stderr)
         print(file=sys.stderr)
 
-        result = client.fetch_artifacts(identifiers=identifiers)
+        result = client.fetch_artifacts(identifiers=identifiers, data_source=data_source)
 
-        print(format_artifacts(result))
+        print(format_artifacts(result, data_source=data_source))
 
     except Exception as e:
         print(f"❌ Error: {e}", file=sys.stderr)
diff --git a/skills/codealive-context-engine/scripts/grep.py b/skills/codealive-context-engine/scripts/grep.py
index f53aa59..2982d85 100755
--- a/skills/codealive-context-engine/scripts/grep.py
+++ b/skills/codealive-context-engine/scripts/grep.py
@@ -42,6 +42,24 @@ def format_grep_results(results: dict) -> str:
             output.append(f"  File: {file_path}")
         if result.get("identifier"):
             output.append(f"  Identifier: {result['identifier']}")
+
+        # Surface the data-source name/id so they can be passed back as --data-source to
+        # fetch.py / relationships.py when an identifier is branch-ambiguous.
+        # dataSource may be a {name, id} object or a bare string, depending on the API response
+        # shape — handle both, mirroring search.py.
+        ds = result.get("dataSource")
+        if isinstance(ds, dict):
+            ds_name = ds.get("name")
+            ds_id = ds.get("id")
+        else:
+            ds_name = ds
+            ds_id = None
+        if ds_name and ds_id:
+            output.append(f"  Source: {ds_name} (id: {ds_id})")
+        elif ds_name:
+            output.append(f"  Source: {ds_name}")
+        elif ds_id:
+            output.append(f"  Source: (id: {ds_id})")
         if result.get("matchCount") is not None:
             output.append(f"  Match count: {result['matchCount']}")
 
diff --git a/skills/codealive-context-engine/scripts/lib/api_client.py b/skills/codealive-context-engine/scripts/lib/api_client.py
index 705481b..8ebda86 100644
--- a/skills/codealive-context-engine/scripts/lib/api_client.py
+++ b/skills/codealive-context-engine/scripts/lib/api_client.py
@@ -524,6 +524,7 @@ def grep_search(
     def fetch_artifacts(
         self,
         identifiers: List[str],
+        data_source: Optional[str] = None,
     ) -> Dict[str, Any]:
         """
         Retrieve full content for code artifacts by their identifiers.
@@ -536,6 +537,10 @@ def fetch_artifacts(
 
         Args:
             identifiers: List of artifact identifiers from search results (max 20)
+            data_source: Optional data-source Name or Id to disambiguate an identifier that
+                exists in more than one data source. Copy the `dataSource.name`/`dataSource.id`
+                from a search result. Omit for normal lookups; an ambiguous identifier without
+                it returns a 409 listing the candidate data sources.
 
         Returns:
             Dict with 'artifacts' list. Each artifact has identifier, content,
@@ -545,6 +550,8 @@ def fetch_artifacts(
             the full list and other relationship profiles.
         """
         body: Dict[str, Any] = {"identifiers": identifiers}
+        if data_source:
+            body["dataSource"] = data_source
         return self._make_request("POST", "/api/search/artifacts", body=body)
 
     def get_artifact_relationships(
@@ -552,6 +559,7 @@ def get_artifact_relationships(
         identifier: str,
         profile: str = "callsOnly",
         max_count_per_type: int = 50,
+        data_source: Optional[str] = None,
     ) -> Dict[str, Any]:
         """
         Retrieve relationship groups for a single artifact by profile.
@@ -569,6 +577,9 @@ def get_artifact_relationships(
                      - "referencesOnly": symbol references
             max_count_per_type: Max related artifacts per relationship type
                                 (1–1000, default 50).
+            data_source: Optional data-source Name or Id to disambiguate a source identifier
+                that exists in more than one data source. Omit for normal lookups; an ambiguous
+                identifier without it returns a 409 listing the candidate data sources.
 
         Returns:
             Dict with sourceIdentifier, profile, found, and a list of
@@ -594,6 +605,8 @@ def get_artifact_relationships(
             "profile": api_profile,
             "maxCountPerType": max_count_per_type,
         }
+        if data_source:
+            body["dataSource"] = data_source
         return self._make_request(
             "POST", "/api/search/artifact-relationships", body=body
         )
@@ -665,8 +678,8 @@ def main():
         print("  search <query> <data_source1> [data_source2...] [--mode auto|fast|deep] [--description-detail short|full]")
         print("  semantic-search <query> <data_source1> [data_source2...] [--path PATH] [--ext EXT] [--max-results N]")
         print("  grep-search <query> <data_source1> [data_source2...] [--regex] [--path PATH] [--ext EXT] [--max-results N]")
-        print("  fetch <identifier1> [identifier2...]")
-        print("  relationships <identifier> [--profile callsOnly|inheritanceOnly|allRelevant|referencesOnly] [--max-count N]")
+        print("  fetch <identifier1> [identifier2...] [--data-source NAME_OR_ID]")
+        print("  relationships <identifier> [--profile callsOnly|inheritanceOnly|allRelevant|referencesOnly] [--max-count N] [--data-source NAME_OR_ID]")
         print("  chat <question> <data_source1> [data_source2...] [--conversation-id ID]")
         sys.exit(1)
 
@@ -791,12 +804,27 @@ def main():
 
         elif command == "fetch":
             if len(sys.argv) < 3:
-                print("Usage: fetch <identifier1> [identifier2...]")
+                print("Usage: fetch <identifier1> [identifier2...] [--data-source NAME_OR_ID]")
                 sys.exit(1)
 
-            identifiers = sys.argv[2:]
+            identifiers = []
+            data_source = None
+            i = 2
+            while i < len(sys.argv):
+                arg = sys.argv[i]
+                if arg == "--data-source":
+                    # Match the flag first, then require a value — otherwise a trailing
+                    # "--data-source" with no value would be silently appended as an identifier.
+                    if i + 1 >= len(sys.argv):
+                        print("Error: --data-source requires a value.", file=sys.stderr)
+                        sys.exit(1)
+                    data_source = sys.argv[i + 1]
+                    i += 2
+                else:
+                    identifiers.append(arg)
+                    i += 1
 
-            result = client.fetch_artifacts(identifiers)
+            result = client.fetch_artifacts(identifiers, data_source=data_source)
             print(json.dumps(result, indent=2))
 
         elif command == "relationships":
@@ -807,20 +835,37 @@ def main():
             identifier = sys.argv[2]
             profile = "callsOnly"
             max_count = 50
+            data_source = None
 
             i = 3
             while i < len(sys.argv):
                 arg = sys.argv[i]
-                if arg == "--profile" and i + 1 < len(sys.argv):
+                # Value-bearing flags match on the name first, then require a value, so a trailing
+                # flag with no value reports "requires a value" instead of being silently skipped.
+                if arg == "--profile":
+                    if i + 1 >= len(sys.argv):
+                        print("Error: --profile requires a value.", file=sys.stderr)
+                        sys.exit(1)
                     profile = sys.argv[i + 1]
                     i += 2
-                elif arg == "--max-count" and i + 1 < len(sys.argv):
+                elif arg == "--max-count":
+                    if i + 1 >= len(sys.argv):
+                        print("Error: --max-count requires a value.", file=sys.stderr)
+                        sys.exit(1)
                     max_count = int(sys.argv[i + 1])
                     i += 2
+                elif arg == "--data-source":
+                    if i + 1 >= len(sys.argv):
+                        print("Error: --data-source requires a value.", file=sys.stderr)
+                        sys.exit(1)
+                    data_source = sys.argv[i + 1]
+                    i += 2
                 else:
                     i += 1
 
-            result = client.get_artifact_relationships(identifier, profile, max_count)
+            result = client.get_artifact_relationships(
+                identifier, profile, max_count, data_source=data_source
+            )
             print(json.dumps(result, indent=2))
 
         elif command == "chat":
@@ -835,7 +880,12 @@ def main():
             i = 3
             while i < len(sys.argv):
                 arg = sys.argv[i]
-                if arg == "--conversation-id" and i + 1 < len(sys.argv):
+                if arg == "--conversation-id":
+                    # Match the flag first, then require a value — otherwise a trailing
+                    # "--conversation-id" with no value would be silently appended as a data source.
+                    if i + 1 >= len(sys.argv):
+                        print("Error: --conversation-id requires a value.", file=sys.stderr)
+                        sys.exit(1)
                     conversation_id = sys.argv[i + 1]
                     i += 2
                 else:
diff --git a/skills/codealive-context-engine/scripts/relationships.py b/skills/codealive-context-engine/scripts/relationships.py
index 5e417ab..273c970 100755
--- a/skills/codealive-context-engine/scripts/relationships.py
+++ b/skills/codealive-context-engine/scripts/relationships.py
@@ -11,7 +11,11 @@
 this script gives you the full list and lets you switch profiles.
 
 Usage:
-    python relationships.py <identifier> [--profile PROFILE] [--max-count N]
+    python relationships.py <identifier> [--profile PROFILE] [--max-count N] [--data-source NAME_OR_ID]
+
+Pass --data-source (a data source Name or Id from a search result's `dataSource`)
+to disambiguate an identifier that exists in more than one data source. Without it,
+an ambiguous identifier returns a 409 listing the candidate data sources.
 
 Profiles:
     callsOnly         (default) outgoing + incoming calls
@@ -61,7 +65,7 @@
 }
 
 
-def format_relationships(data: dict) -> str:
+def format_relationships(data: dict, data_source: str = None) -> str:
     """Format an artifact-relationships response for display."""
     source_id = data.get("sourceIdentifier") or "<unknown>"
     raw_profile = data.get("profile") or ""
@@ -69,10 +73,19 @@ def format_relationships(data: dict) -> str:
     found = bool(data.get("found"))
 
     if not found:
-        return (
-            f"Artifact not found or inaccessible: {source_id}\n"
-            f"(profile={profile})"
-        )
+        lines = [
+            f"Artifact not found or inaccessible: {source_id}",
+            f"(profile={profile})",
+        ]
+        if data_source:
+            lines.append(
+                f'\n💡 Hint: nothing matched in data source "{data_source}". The identifier may belong '
+                "to a different data source, or the --data-source value may be wrong. Try: re-run with "
+                "--data-source set to a different candidate (use the Source name or id from your "
+                "search results, or run datasources.py), or drop --data-source entirely — an ambiguous "
+                "identifier then returns a 409 listing the candidate data sources to choose from."
+            )
+        return "\n".join(lines)
 
     relationships = data.get("relationships") or []
 
@@ -142,20 +155,35 @@ def main():
     identifier = sys.argv[1]
     profile = "callsOnly"
     max_count = 50
+    data_source = None
 
     i = 2
     while i < len(sys.argv):
         arg = sys.argv[i]
-        if arg == "--profile" and i + 1 < len(sys.argv):
+        # Value-bearing flags match on the name first, then require a value, so a trailing flag with
+        # no value reports "requires a value" rather than the misleading "unknown argument" below.
+        if arg == "--profile":
+            if i + 1 >= len(sys.argv):
+                print("Error: --profile requires a value.", file=sys.stderr)
+                sys.exit(1)
             profile = sys.argv[i + 1]
             i += 2
-        elif arg == "--max-count" and i + 1 < len(sys.argv):
+        elif arg == "--max-count":
+            if i + 1 >= len(sys.argv):
+                print("Error: --max-count requires a value.", file=sys.stderr)
+                sys.exit(1)
             try:
                 max_count = int(sys.argv[i + 1])
             except ValueError:
                 print(f"Error: --max-count expects an integer, got '{sys.argv[i + 1]}'", file=sys.stderr)
                 sys.exit(1)
             i += 2
+        elif arg == "--data-source":
+            if i + 1 >= len(sys.argv):
+                print("Error: --data-source requires a value.", file=sys.stderr)
+                sys.exit(1)
+            data_source = sys.argv[i + 1]
+            i += 2
         elif arg == "--json":
             # Handled below — we strip it before calling format_relationships
             i += 1
@@ -171,18 +199,21 @@ def main():
 
         print(f"🔗 Fetching {profile} relationships for: {identifier}", file=sys.stderr)
         print(f"⚙️  max-count={max_count}", file=sys.stderr)
+        if data_source:
+            print(f"   data source: {data_source}", file=sys.stderr)
         print(file=sys.stderr)
 
         result = client.get_artifact_relationships(
             identifier=identifier,
             profile=profile,
             max_count_per_type=max_count,
+            data_source=data_source,
         )
 
         if as_json:
             print(json.dumps(result, indent=2))
         else:
-            print(format_relationships(result))
+            print(format_relationships(result, data_source=data_source))
 
     except Exception as e:
         print(f"❌ Error: {e}", file=sys.stderr)
diff --git a/skills/codealive-context-engine/scripts/search.py b/skills/codealive-context-engine/scripts/search.py
index 195fae1..99a21be 100755
--- a/skills/codealive-context-engine/scripts/search.py
+++ b/skills/codealive-context-engine/scripts/search.py
@@ -57,7 +57,12 @@ def format_search_results(results: dict) -> str:
         end_line = range_info.get("end", {}).get("line") or result.get("endLine")
 
         ds = result.get("dataSource", {})
-        source_name = ds.get("name") if isinstance(ds, dict) else ds
+        if isinstance(ds, dict):
+            source_name = ds.get("name")
+            source_id = ds.get("id")
+        else:
+            source_name = ds
+            source_id = None
 
         kind = result.get("kind", "")
         identifier = result.get("identifier", "")
@@ -84,8 +89,14 @@ def format_search_results(results: dict) -> str:
             short_id = identifier.split("::")[-1] if "::" in identifier else identifier
             if short_id != file_path:
                 output.append(f"  Symbol: {short_id}")
-        if source_name:
+        # Surface both the data-source name and id so they can be passed straight back as
+        # --data-source to fetch.py / relationships.py when an identifier is branch-ambiguous.
+        if source_name and source_id:
+            output.append(f"  Source: {source_name} (id: {source_id})")
+        elif source_name:
             output.append(f"  Source: {source_name}")
+        elif source_id:
+            output.append(f"  Source: (id: {source_id})")
         if content_byte_size is not None:
             output.append(f"  Size: {_format_byte_size(content_byte_size)}")
         if description:
diff --git a/tests/test_cli_smoke.py b/tests/test_cli_smoke.py
index 5625c16..00492ad 100644
--- a/tests/test_cli_smoke.py
+++ b/tests/test_cli_smoke.py
@@ -309,3 +309,84 @@ def test_check_auth_hook_normalizes_base_url_and_uses_repo_root_fallback():
     assert result.returncode == 0
     assert "https://codealive.example.com/settings/api-keys" in result.stdout
     assert str(REPO_ROOT / "skills" / "codealive-context-engine" / "setup.py") in result.stdout
+
+
+def test_fetch_and_relationships_scripts_hint_when_data_source_misses():
+    """A --data-source-scoped request that finds nothing must hint: try another source or drop it."""
+    def fetch_handler(request):
+        body = json.loads(request["body"])
+        assert body["dataSource"] == "backend"
+        # Scoped to a source that doesn't contain the identifier -> null content.
+        return 200, {"artifacts": [{"identifier": body["identifiers"][0], "content": None}]}, {}
+
+    def relationships_handler(request):
+        body = json.loads(request["body"])
+        assert body["dataSource"] == "backend"
+        return 200, {
+            "sourceIdentifier": body["identifier"],
+            "profile": body["profile"],
+            "found": False,
+            "relationships": [],
+        }, {}
+
+    with mock_codealive_server(
+        {
+            ("POST", "/api/search/artifacts"): fetch_handler,
+            ("POST", "/api/search/artifact-relationships"): relationships_handler,
+        }
+    ) as (base_url, _requests):
+        env = {
+            **os.environ,
+            "CODEALIVE_API_KEY": "skill-test-key",
+            "CODEALIVE_BASE_URL": f"{base_url}/api",
+        }
+
+        fetch = _run(
+            "fetch.py", "org/repo::src/auth.py::Missing", "--data-source", "backend", env=env
+        )
+        rel = _run(
+            "relationships.py", "org/repo::src/auth.py::Missing", "--data-source", "backend", env=env
+        )
+
+    assert fetch.returncode == 0, fetch.stderr
+    assert "backend" in fetch.stdout
+    assert "--data-source" in fetch.stdout
+    assert "drop --data-source" in fetch.stdout
+
+    assert rel.returncode == 0, rel.stderr
+    assert "backend" in rel.stdout
+    assert "drop --data-source" in rel.stdout
+
+
+def test_fetch_and_relationships_scripts_error_on_data_source_without_value():
+    """A trailing --data-source with no value must error, not be silently treated as an identifier."""
+    env = {**os.environ, "CODEALIVE_API_KEY": "skill-test-key"}
+
+    fetch = _run("fetch.py", "org/repo::src/auth.py::AuthService", "--data-source", env=env)
+    assert fetch.returncode != 0
+    assert "--data-source requires a value" in fetch.stderr
+
+    rel = _run("relationships.py", "org/repo::src/auth.py::AuthService", "--data-source", env=env)
+    assert rel.returncode != 0
+    assert "--data-source requires a value" in rel.stderr
+
+
+def test_grep_format_surfaces_dict_and_string_datasource():
+    """format_grep_results surfaces the source whether dataSource is a {name,id} object or a string."""
+    import importlib.util
+
+    grep_path = SKILL_ROOT / "scripts" / "grep.py"
+    spec = importlib.util.spec_from_file_location("grep_under_test", grep_path)
+    grep_mod = importlib.util.module_from_spec(spec)
+    spec.loader.exec_module(grep_mod)
+
+    dict_out = grep_mod.format_grep_results(
+        {"results": [{"identifier": "org/repo::a.py::F", "dataSource": {"name": "backend", "id": "abc123"}}]}
+    )
+    assert "Source: backend (id: abc123)" in dict_out
+
+    # A bare-string dataSource must still be surfaced (was silently dropped before).
+    str_out = grep_mod.format_grep_results(
+        {"results": [{"identifier": "org/repo::a.py::F", "dataSource": "backend"}]}
+    )
+    assert "Source: backend" in str_out
diff --git a/tests/test_setup_and_client.py b/tests/test_setup_and_client.py
index a11c1b3..fbc507b 100644
--- a/tests/test_setup_and_client.py
+++ b/tests/test_setup_and_client.py
@@ -325,6 +325,70 @@ def test_api_client_get_artifact_relationships_rejects_unknown_profile():
         raise AssertionError("ValueError was not raised for unknown profile")
 
 
+def test_api_client_fetch_artifacts_forwards_data_source():
+    received_bodies: list = []
+
+    def fetch_handler(request):
+        body = json.loads(request["body"])
+        received_bodies.append(body)
+        return 200, {"artifacts": []}, {}
+
+    with mock_codealive_server(
+        {("POST", "/api/search/artifacts"): fetch_handler}
+    ) as (base_url, _requests):
+        client = CodeAliveClient(api_key="skill-test-key", base_url=base_url)
+        # Omitted by default…
+        client.fetch_artifacts(["org/repo::src/a.py::F"])
+        # …forwarded as DataSource when provided.
+        client.fetch_artifacts(["org/repo::src/a.py::F"], data_source="backend")
+
+    assert "dataSource" not in received_bodies[0]
+    assert received_bodies[1]["dataSource"] == "backend"
+
+
+def test_api_client_get_artifact_relationships_forwards_data_source():
+    received_bodies: list = []
+
+    def relationships_handler(request):
+        body = json.loads(request["body"])
+        received_bodies.append(body)
+        return 200, {"sourceIdentifier": body["identifier"], "profile": body["profile"], "found": True, "relationships": []}, {}
+
+    with mock_codealive_server(
+        {("POST", "/api/search/artifact-relationships"): relationships_handler}
+    ) as (base_url, _requests):
+        client = CodeAliveClient(api_key="skill-test-key", base_url=base_url)
+        client.get_artifact_relationships("org/repo::src/a.py::F")
+        client.get_artifact_relationships("org/repo::src/a.py::F", data_source="ds-main")
+
+    assert "dataSource" not in received_bodies[0]
+    assert received_bodies[1]["dataSource"] == "ds-main"
+
+
+def test_api_client_ambiguous_409_surfaces_candidate_data_sources():
+    # When an identifier is ambiguous and no data_source is supplied, the backend returns a 409
+    # whose detail lists the candidate data sources. The client must surface those candidates so
+    # the agent can retry with --data-source rather than inventing a result.
+    def fetch_handler(request):
+        return 409, {
+            "title": "Ambiguous data source",
+            "detail": "Identifier matches 2 data sources: Name='backend' Id='ds-main', Name='backend-legacy' Id='ds-master'",
+        }, {}
+
+    with mock_codealive_server(
+        {("POST", "/api/search/artifacts"): fetch_handler}
+    ) as (base_url, _requests):
+        client = CodeAliveClient(api_key="skill-test-key", base_url=base_url)
+        try:
+            client.fetch_artifacts(["org/repo::src/a.py::F"])
+        except Exception as e:
+            message = str(e)
+            assert "409" in message
+            assert "backend" in message and "backend-legacy" in message
+        else:
+            raise AssertionError("Expected an exception for the ambiguous 409 response")
+
+
 # ===== Phase 1 — error contract & ObjectId preflight =====
 
 def test_format_codealive_error_renders_rfc9457_problem_details():