From 704bbcca804c443dbc0d0ad82d7fc50b6549380d Mon Sep 17 00:00:00 2001
From: Teknium <127238744+teknium1@users.noreply.github.com>
Date: Wed, 15 Jul 2026 07:42:33 -0700
Subject: [PATCH] feat(compressor): keep image URLs and unknown-part markers in
 summarizer serialization
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Follow-up on @AlexFucuson9's cherry-picked multimodal flattening:

- http(s) image parts render as '[image: <url>]' so the summary keeps a
  referenceable handle after compaction (base64 data: URLs still
  collapse to '[image]' — no reusable reference, and leaking them is
  the bug being fixed)
- unknown part types (document, future shapes) render as '[<type>]'
  instead of being silently dropped
- test docstrings corrected: the original crash premise is stale
  (redact_sensitive_text grew str() coercion in #52147); the live bug
  is repr-noise + base64 leakage into the summarizer input
---
 agent/context_compressor.py                   | 34 +++++++++++++---
 .../agent/test_compressor_media_stripping.py  | 40 ++++++++++++++++---
 2 files changed, 63 insertions(+), 11 deletions(-)

diff --git a/agent/context_compressor.py b/agent/context_compressor.py
index e16778cd3..b652d4ceb 100644
--- a/agent/context_compressor.py
+++ b/agent/context_compressor.py
@@ -586,6 +586,27 @@ def _strip_historical_media(messages: List[Dict[str, Any]]) -> List[Dict[str, An
     return result if changed else messages
 
 
+def _image_part_label(part: Dict[str, Any]) -> str:
+    """Render a multimodal image part as a short text label for the summarizer.
+
+    Keeps a real, referenceable URL when the image lives at an http(s)
+    address — the summary can then preserve the handle so the agent (or a
+    later vision_analyze call) can still reach the image after compaction.
+    Base64 ``data:`` URLs carry no reusable reference and would flood the
+    summarizer input, so they collapse to ``[image]``.
+    """
+    url = ""
+    if isinstance(part.get("image_url"), dict):
+        url = str(part["image_url"].get("url") or "")
+    elif isinstance(part.get("image_url"), str):
+        url = part["image_url"]
+    elif isinstance(part.get("url"), str):
+        url = part["url"]
+    if url.startswith(("http://", "https://")):
+        return f"[image: {url}]"
+    return "[image]"
+
+
 def _str_arg(args: dict, key: str, default: str = "") -> str:
     """Safely get a string argument from parsed tool args.
 
@@ -1669,12 +1690,15 @@ class ContextCompressor(ContextEngine):
                 text_parts: list[str] = []
                 for part in content:
                     if isinstance(part, dict):
-                        if part.get("type") == "text":
+                        ptype = part.get("type")
+                        if ptype == "text":
                             text_parts.append(part.get("text", ""))
-                        elif part.get("type") in {
-                            "image", "image_url", "input_image",
-                        }:
-                            text_parts.append("[image]")
+                        elif ptype in {"image", "image_url", "input_image"}:
+                            text_parts.append(_image_part_label(part))
+                        else:
+                            # Unknown part type — keep a marker so the
+                            # summarizer knows content existed here.
+                            text_parts.append(f"[{ptype or 'attachment'}]")
                     elif isinstance(part, str):
                         text_parts.append(part)
                 content = "\n".join(text_parts)
diff --git a/tests/agent/test_compressor_media_stripping.py b/tests/agent/test_compressor_media_stripping.py
index 29ca272a4..2ae9cbf59 100644
--- a/tests/agent/test_compressor_media_stripping.py
+++ b/tests/agent/test_compressor_media_stripping.py
@@ -56,12 +56,11 @@ class TestMediaDirectiveStripping:
         assert result.count("[media attachment]") == 2
 
     def test_multimodal_list_content_does_not_crash(self, compressor):
-        """content as a list (multimodal) must not crash redact_sensitive_text.
+        """content as a list (multimodal) must be flattened to clean text.
 
-        Regression: msg.get('content') can return a list of parts for
-        multimodal messages (images + text).  The old code passed this
-        list directly to redact_sensitive_text(str), which raised
-        AttributeError on list.replace().
+        Without flattening, str() coercion in redact_sensitive_text dumps
+        the raw part-dict repr — including base64 image data — into the
+        summarizer input, burning summary budget on noise.
         """
         turns = [
             {
@@ -72,10 +71,39 @@ class TestMediaDirectiveStripping:
                 ],
             },
         ]
-        # Must not raise
         result = compressor._serialize_for_summary(turns)
         assert "What is in this image?" in result
         assert "[image]" in result
+        assert "base64" not in result
+
+    def test_multimodal_remote_image_keeps_url(self, compressor):
+        """http(s) image parts keep their URL as a referenceable handle."""
+        turns = [
+            {
+                "role": "user",
+                "content": [
+                    {"type": "text", "text": "look at this"},
+                    {"type": "image_url", "image_url": {"url": "https://example.com/a.png"}},
+                ],
+            },
+        ]
+        result = compressor._serialize_for_summary(turns)
+        assert "[image: https://example.com/a.png]" in result
+
+    def test_multimodal_unknown_part_type_keeps_marker(self, compressor):
+        """Unknown part types are not silently dropped."""
+        turns = [
+            {
+                "role": "user",
+                "content": [
+                    {"type": "text", "text": "see attachment"},
+                    {"type": "document", "title": "spec.pdf"},
+                ],
+            },
+        ]
+        result = compressor._serialize_for_summary(turns)
+        assert "see attachment" in result
+        assert "[document]" in result
 
     def test_multimodal_list_text_parts_extracted(self, compressor):
         """Text parts from multimodal list content are preserved in output."""