Reply sanitizer: cap raw input, auto-close open tags (no severed HTML)

Per Codex: slicing the SANITIZED html with [:8000] could cut through a tag or
entity. Cap the RAW editor HTML (20k) before sanitizing instead, and have
sanitize_reply_html auto-close any still-open allowed tags so malformed input
can never leave a dangling/severed tag in message_html or the email body.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

goodnews/api.py

@@ -843,8 +843,9 @@ def create_app() -> FastAPI:
 
     @app.post("/api/admin/feedback/{fid}/reply")
     def admin_feedback_reply(fid: int, body: FeedbackReplyBody, request: Request) -> dict:
-        # Sanitize the editor HTML to our allowlist; derive the plain-text fallback.
-        reply_html = sanitize_reply_html(body.html)[:8000]
+        # Cap the RAW editor HTML first (slicing sanitized output could sever a
+        # tag), then sanitize the whole thing.
+        reply_html = sanitize_reply_html((body.html or "")[:20000])
         reply_text = reply_html_to_text(reply_html)
         if not reply_text:
             raise HTTPException(status_code=422, detail="Reply message is required.")

goodnews/markup.py

@@ -118,6 +118,11 @@ def sanitize_reply_html(raw: str) -> str:
     p = _Sanitizer()
     p.feed(raw)
     p.close()
+    # Close any still-open allowed tags (malformed input → never emit a dangling
+    # or severed tag into stored HTML / the email body).
+    for canon in reversed(p.open):
+        if canon:
+            p.out.append(f"</{canon}>")
     html = "".join(p.out)
     # If nothing but markup/whitespace survived, treat as empty.
     if not re.sub(r"<[^>]+>", "", html).strip():