From 21b078271ab64fdb9bac268f312c558cc78c0d15 Mon Sep 17 00:00:00 2001 From: Julia Pagnucco Date: Mon, 16 Nov 2026 16:03:48 +0000 Subject: [PATCH] Add non-normative soft maximum field length guidance --- index.bs | 95 +++++++++++++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 91 insertions(+), 4 deletions(-) diff --git a/index.bs b/index.bs index bbb561c..f629ef4 100644 --- a/index.bs +++ b/index.bs @@ -1718,7 +1718,9 @@ Prompt injection attacks in WebMCP can be categorized by three distinct attack v
Metadata / Description Attacks (Tool Poisoning)
-Malicious instructions embedded in tool metadata (name, description, parameter descriptions) that manipulate agent behavior. +Malicious instructions embedded in tool metadata ({{ModelContextTool/name}}, +{{ModelContextTool/title}}, {{ModelContextTool/description}}, {{ModelContextTool/inputSchema}}, and +{{ModelContextTool/outputSchema}}) that manipulate agent behavior.