feat: best-effort compact large tool schemas (#23904)

## Why

The `dev/cc/ref-def` branch preserves richer JSON Schema detail for
connector tools, including `$defs` and nested shapes. That improves
fidelity, but it pushes the largest connector schemas well past the
intended tool-schema budget. This PR adds a best-effort compaction pass
for unusually large tool input schemas so the p99 and max tails stay
small while ordinary schemas are left alone.

## What Changed

- Added best-effort large-schema compaction in
`codex-rs/tools/src/json_schema.rs` after schema sanitization and
definition pruning.
- Compaction runs as a waterfall only while the compact JSON budget
proxy is exceeded:
  1. Strip schema `description` metadata.
  2. Drop root `$defs` / `definitions`.
  3. Collapse deep nested complex schema objects to `{}`.
- Kept top-level argument names and immediate schema shape where
possible.

## Corpus Results

Scope: 2,025 schemas under `golden_schemas`, all parsed successfully.
Token count is `o200k_base` over compact JSON from
`parse_tool_input_schema`.

| Percentile | Before `origin/main` `4dbca61e20` | After branch
`dev/cc/ref-def` `f9bf071758` | After this PR |
|---|---:|---:|---:|
| p0 | 9 | 9 | 9 |
| p10 | 59 | 63 | 63 |
| p25 | 81 | 86 | 86 |
| p50 | 114 | 127 | 125 |
| p75 | 174 | 205 | 202 |
| p90 | 295 | 335 | 322 |
| p95 | 391 | 526 | 422 |
| p99 | 794 | 1,303 | 689 |
| max | 2,836 | 3,337 | 887 |

After this PR, `0 / 2,025` schemas are over 1k tokens.

### Compaction Savings

These are cumulative waterfall stages over the same corpus. Later passes
only run for schemas that are still over the compact JSON budget proxy.

| Stage | Total tokens | Step savings | Schemas changed by step |
|---|---:|---:|---:|
| No compaction | 391,862 | - | - |
| Strip schema `description` metadata | 350,961 | 40,901 | 66 |
| Drop root `$defs` / `definitions` | 340,683 | 10,278 | 13 |
| Collapse deep complex schemas to `{}` | 335,875 | 4,808 | 6 |

codex-rs/tools/src/json_schema.rs

@@ -160,6 +160,7 @@ pub fn parse_tool_input_schema(input_schema: &JsonValue) -> Result<JsonSchema, s
     let mut input_schema = input_schema.clone();
     sanitize_json_schema(&mut input_schema);
     prune_unreachable_definitions(&mut input_schema);
+    compact_large_tool_schema(&mut input_schema);
     let schema: JsonSchema = serde_json::from_value(input_schema)?;
     if matches!(
         schema.schema_type,
@@ -170,6 +171,47 @@ pub fn parse_tool_input_schema(input_schema: &JsonValue) -> Result<JsonSchema, s
     Ok(schema)
 }
 
+// Use compact normalized JSON bytes as a cheap local proxy for the 1k-token
+// schema budget.
+const MAX_COMPACT_TOOL_SCHEMA_BYTES: usize = 4_000;
+const MAX_COMPACT_TOOL_SCHEMA_DEPTH: usize = 2;
+
+/// Shrink unusually large tool schemas while preserving the top-level argument
+/// surface. Compaction is best-effort rather than a hard cap: it runs only
+/// after schema sanitization/pruning and applies increasingly lossy passes
+/// while the schema remains over budget.
+fn compact_large_tool_schema(value: &mut JsonValue) {
+    for pass in LARGE_SCHEMA_COMPACTION_PASSES {
+        if compact_schema_fits_budget(value) {
+            break;
+        }
+        pass(value);
+    }
+}
+
+type LargeSchemaCompactionPass = fn(&mut JsonValue);
+
+const LARGE_SCHEMA_COMPACTION_PASSES: &[LargeSchemaCompactionPass] = &[
+    strip_schema_descriptions,
+    drop_schema_definitions,
+    collapse_deep_schema_objects_from_root,
+];
+
+fn collapse_deep_schema_objects_from_root(value: &mut JsonValue) {
+    collapse_deep_schema_objects(value, /*depth*/ 0);
+}
+
+fn compact_schema_fits_budget(value: &JsonValue) -> bool {
+    compact_normalized_schema_len(value) <= MAX_COMPACT_TOOL_SCHEMA_BYTES
+}
+
+fn compact_normalized_schema_len(value: &JsonValue) -> usize {
+    serde_json::from_value::<JsonSchema>(value.clone())
+        .and_then(|schema| serde_json::to_vec(&schema))
+        .map(|json| json.len())
+        .unwrap_or(0)
+}
+
 #[derive(Debug, Clone, Copy, PartialEq, Eq)]
 enum DefinitionTraversal {
     Include,
@@ -214,6 +256,130 @@ fn for_each_schema_child(
     }
 }
 
+fn strip_schema_descriptions(value: &mut JsonValue) {
+    match value {
+        JsonValue::Array(values) => {
+            for value in values {
+                strip_schema_descriptions(value);
+            }
+        }
+        JsonValue::Object(map) => {
+            map.remove("description");
+            for_each_schema_child_mut(map, DefinitionTraversal::Include, &mut |value| {
+                strip_schema_descriptions(value);
+            });
+        }
+        _ => {}
+    }
+}
+
+fn for_each_schema_child_mut(
+    map: &mut serde_json::Map<String, JsonValue>,
+    definition_traversal: DefinitionTraversal,
+    visitor: &mut impl FnMut(&mut JsonValue),
+) {
+    if let Some(properties) = map.get_mut("properties")
+        && let Some(properties_map) = properties.as_object_mut()
+    {
+        for value in properties_map.values_mut() {
+            visitor(value);
+        }
+    }
+
+    for key in SCHEMA_CHILD_KEYS {
+        if let Some(value) = map.get_mut(key) {
+            visitor(value);
+        }
+    }
+
+    if let Some(additional_properties) = map.get_mut("additionalProperties")
+        && !matches!(additional_properties, JsonValue::Bool(_))
+    {
+        visitor(additional_properties);
+    }
+
+    if definition_traversal == DefinitionTraversal::Include {
+        for key in DEFINITION_TABLE_KEYS {
+            if let Some(definitions) = map.get_mut(key)
+                && let Some(definitions_map) = definitions.as_object_mut()
+            {
+                for value in definitions_map.values_mut() {
+                    visitor(value);
+                }
+            }
+        }
+    }
+}
+
+/// Replace local definition refs with empty schemas before dropping root
+/// definition tables, so downstream behavior does not depend on how a schema
+/// parser handles refs to missing definitions.
+fn drop_schema_definitions(value: &mut JsonValue) {
+    rewrite_definition_refs_to_empty_schemas(value);
+
+    let JsonValue::Object(map) = value else {
+        return;
+    };
+
+    for key in DEFINITION_TABLE_KEYS {
+        map.remove(key);
+    }
+}
+
+fn rewrite_definition_refs_to_empty_schemas(value: &mut JsonValue) {
+    match value {
+        JsonValue::Array(values) => {
+            for value in values {
+                rewrite_definition_refs_to_empty_schemas(value);
+            }
+        }
+        JsonValue::Object(map) => {
+            if map
+                .get("$ref")
+                .and_then(JsonValue::as_str)
+                .and_then(parse_local_definition_ref)
+                .is_some()
+            {
+                *value = json!({});
+                return;
+            }
+
+            for_each_schema_child_mut(map, DefinitionTraversal::Skip, &mut |value| {
+                rewrite_definition_refs_to_empty_schemas(value);
+            });
+        }
+        _ => {}
+    }
+}
+
+fn collapse_deep_schema_objects(value: &mut JsonValue, depth: usize) {
+    match value {
+        JsonValue::Array(values) => {
+            for value in values {
+                collapse_deep_schema_objects(value, depth);
+            }
+        }
+        JsonValue::Object(map) => {
+            if depth >= MAX_COMPACT_TOOL_SCHEMA_DEPTH && is_complex_schema_object(map) {
+                *value = json!({});
+                return;
+            }
+
+            for_each_schema_child_mut(map, DefinitionTraversal::Skip, &mut |value| {
+                collapse_deep_schema_objects(value, depth + 1);
+            });
+        }
+        _ => {}
+    }
+}
+
+fn is_complex_schema_object(map: &serde_json::Map<String, JsonValue>) -> bool {
+    SCHEMA_CHILD_KEYS.iter().any(|key| map.contains_key(*key))
+        || map.contains_key("properties")
+        || map.contains_key("additionalProperties")
+        || map.contains_key("$ref")
+}
+
 /// Sanitize a JSON Schema (as serde_json::Value) so it can fit our limited
 /// schema representation. This function:
 /// - Ensures every typed schema object has a `"type"` when required.