fix(agent-tools): render tool schemas without HTML-safe escapes

minijinja's tojson escapes `<`, `>`, `&` and `'` so its output can ride
a script tag. A system prompt is not a script tag, so every apostrophe
in a field description reached the model as \u0027 and every
`<provider>` as \u003c: harder to read, and more tokens for the same
text.

The renderer registers a json filter that serializes through serde_json
and leaves those characters alone, and both tool sections use it.
tojson stays for a user template whose output does feed a page. The
serialization is also compact, which drops the spaces tojson puts
after every separator.

The tool row is `name | arguments` with the description indented
underneath, so a multi-line description can no longer push the schema
column below another tool's prose, where it read as part of that tool.

crates/rhema-claw/SYSTEM.md

@@ -19,11 +19,12 @@ it, is the reply the user sees.
 {% if tools %}
 ## Tools
 
-You act through these tools (name | description | arguments). The
-arguments are the JSON schema a call must satisfy.
+You act through these tools: each entry is its name and arguments, then
+what it is for. The arguments are the JSON schema a call must satisfy.
 
 {% for tool in tools %}
-- {{ tool.name }} | {{ tool.description }} | {{ tool.parameters | tojson }}
+- {{ tool.name }} | {{ tool.parameters | json }}
+{{ tool.description | indent(2, true) }}
 {% endfor %}
 {% endif %}
 

crates/rhema-claw/src/bootstrap.rs

@@ -518,7 +518,8 @@ mod tests {
             None,
         )
         .expect("renders");
-        assert!(prompt.contains("- my_tool | its purpose | "), "{prompt}");
+        assert!(prompt.contains("- my_tool | {"), "{prompt}");
+        assert!(prompt.contains("\n  its purpose"), "{prompt}");
         assert!(
             prompt.contains(r#""required":["kind"]"#),
             "the JSON schema rendered: {prompt}"