scriptoriumadmin/fabro
1
0
Fork 0
mirror of https://github.com/fabro-sh/fabro.git synced 2026-10-08 03:10:26 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 5

checkpoint

Browse source 
⚒️ Generated with [Fabro](https://fabro.sh)
This commit is contained in:
Fabro 2026-03-20 18:28:21 -04:00
parent bdcc985dec
commit 07e3306abf
6 changed files with 1093 additions and 21 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

67
checkpoint.json
View file

File diff suppressed because one or more lines are too long

822
nodes/implement/diff.patch Normal file
View file

@ -0,0 +1,822 @@
diff --git a/lib/crates/fabro-agent/src/compaction.rs b/lib/crates/fabro-agent/src/compaction.rs
index 699a6bc5..082c4b9b 100644
--- a/lib/crates/fabro-agent/src/compaction.rs
+++ b/lib/crates/fabro-agent/src/compaction.rs
@@ -112,6 +112,7 @@ function names, error messages, and exact values. Omit pleasantries and conversa
max_tokens: Some(4096),
stop_sequences: None,
reasoning_effort: None,
+ speed: None,
metadata: None,
provider_options: None,
};
diff --git a/lib/crates/fabro-agent/src/config.rs b/lib/crates/fabro-agent/src/config.rs
index 2758dab3..dff155c5 100644
--- a/lib/crates/fabro-agent/src/config.rs
+++ b/lib/crates/fabro-agent/src/config.rs
@@ -64,6 +64,7 @@ pub struct SessionConfig {
pub default_command_timeout_ms: u64,
pub max_command_timeout_ms: u64,
pub reasoning_effort: Option<String>,
+ pub speed: Option<String>,
pub tool_output_limits: HashMap<String, usize>,
pub tool_line_limits: HashMap<String, usize>,
/// Override the provider's default max_tokens when set.
@@ -133,6 +134,7 @@ impl Default for SessionConfig {
max_command_timeout_ms: 600_000,
max_tokens: None,
reasoning_effort: None,
+ speed: None,
tool_output_limits: HashMap::new(),
tool_line_limits: HashMap::new(),
enable_loop_detection: true,
diff --git a/lib/crates/fabro-agent/src/session.rs b/lib/crates/fabro-agent/src/session.rs
index 4fd442dc..dba09be5 100644
--- a/lib/crates/fabro-agent/src/session.rs
+++ b/lib/crates/fabro-agent/src/session.rs
@@ -459,6 +459,10 @@ impl Session {
self.config.reasoning_effort = effort;
}
+ pub fn set_speed(&mut self, speed: Option<String>) {
+ self.config.speed = speed;
+ }
+
pub const fn set_max_turns(&mut self, max_turns: usize) {
self.config.max_turns = max_turns;
}
@@ -906,6 +910,7 @@ impl Session {
}),
stop_sequences: None,
reasoning_effort: self.config.reasoning_effort.clone(),
+ speed: self.config.speed.clone(),
metadata: None,
provider_options: self.provider_profile.provider_options(),
}
diff --git a/lib/crates/fabro-agent/src/tools.rs b/lib/crates/fabro-agent/src/tools.rs
index f204dadb..dc7ed382 100644
--- a/lib/crates/fabro-agent/src/tools.rs
+++ b/lib/crates/fabro-agent/src/tools.rs
@@ -588,6 +588,7 @@ pub(crate) fn make_web_fetch_tool(summarizer: Option<WebFetchSummarizer>) -> Reg
max_tokens: None,
stop_sequences: None,
reasoning_effort: None,
+ speed: None,
metadata: None,
provider_options: None,
};
diff --git a/lib/crates/fabro-agent/src/types.rs b/lib/crates/fabro-agent/src/types.rs
index 8fd178b8..2f7228d7 100644
--- a/lib/crates/fabro-agent/src/types.rs
+++ b/lib/crates/fabro-agent/src/types.rs
@@ -624,6 +624,7 @@ mod tests {
cache_read_tokens: Some(80),
cache_write_tokens: Some(10),
reasoning_tokens: Some(20),
+ speed: None,
raw: None,
};
let event = AgentEvent::AssistantMessage {
diff --git a/lib/crates/fabro-api/src/server.rs b/lib/crates/fabro-api/src/server.rs
index 8da054ac..083fcf67 100644
--- a/lib/crates/fabro-api/src/server.rs
+++ b/lib/crates/fabro-api/src/server.rs
@@ -1232,6 +1232,7 @@ async fn create_completion(
Some(req.stop_sequences)
},
reasoning_effort: req.reasoning_effort,
+ speed: None,
metadata: None,
provider_options: req.provider_options,
};
diff --git a/lib/crates/fabro-cli/src/doctor.rs b/lib/crates/fabro-cli/src/doctor.rs
index 1747f40d..955b6f0b 100644
--- a/lib/crates/fabro-cli/src/doctor.rs
+++ b/lib/crates/fabro-cli/src/doctor.rs
@@ -877,6 +877,7 @@ async fn probe_llm_provider(
max_tokens: Some(16),
stop_sequences: None,
reasoning_effort: None,
+ speed: None,
metadata: None,
provider_options: None,
};
diff --git a/lib/crates/fabro-graphviz/src/graph/types.rs b/lib/crates/fabro-graphviz/src/graph/types.rs
index d6cc8cb6..5da77102 100644
--- a/lib/crates/fabro-graphviz/src/graph/types.rs
+++ b/lib/crates/fabro-graphviz/src/graph/types.rs
@@ -210,6 +210,11 @@ impl Node {
self.str_attr("reasoning_effort").unwrap_or("high")
}
+ #[must_use]
+ pub fn speed(&self) -> Option<&str> {
+ self.str_attr("speed")
+ }
+
#[must_use]
pub fn auto_status(&self) -> bool {
self.bool_attr("auto_status").unwrap_or(false)
@@ -534,6 +539,7 @@ mod tests {
assert_eq!(node.model(), None);
assert_eq!(node.provider(), None);
assert_eq!(node.reasoning_effort(), "high");
+ assert_eq!(node.speed(), None);
assert!(!node.auto_status());
assert!(!node.allow_partial());
assert_eq!(node.retry_policy(), None);
diff --git a/lib/crates/fabro-hooks/src/executor.rs b/lib/crates/fabro-hooks/src/executor.rs
index 4d63e6da..bfb90306 100644
--- a/lib/crates/fabro-hooks/src/executor.rs
+++ b/lib/crates/fabro-hooks/src/executor.rs
@@ -343,6 +343,7 @@ impl HookExecutorImpl {
max_tokens: None,
stop_sequences: None,
reasoning_effort: None,
+ speed: None,
metadata: None,
provider_options: None,
};
diff --git a/lib/crates/fabro-llm/src/client.rs b/lib/crates/fabro-llm/src/client.rs
index 5fb60337..00b81993 100644
--- a/lib/crates/fabro-llm/src/client.rs
+++ b/lib/crates/fabro-llm/src/client.rs
@@ -339,6 +339,7 @@ mod tests {
max_tokens: None,
stop_sequences: None,
reasoning_effort: None,
+ speed: None,
metadata: None,
provider_options: None,
}
diff --git a/lib/crates/fabro-llm/src/generate.rs b/lib/crates/fabro-llm/src/generate.rs
index 807a41f5..815feb79 100644
--- a/lib/crates/fabro-llm/src/generate.rs
+++ b/lib/crates/fabro-llm/src/generate.rs
@@ -69,6 +69,7 @@ fn build_request(
max_tokens: params.max_tokens,
stop_sequences: params.stop_sequences.clone(),
reasoning_effort: params.reasoning_effort.clone(),
+ speed: params.speed.clone(),
metadata: params.metadata.clone(),
provider_options: params.provider_options.clone(),
}
@@ -288,6 +289,7 @@ pub struct GenerateParams {
pub max_tokens: Option<i64>,
pub stop_sequences: Option<Vec<String>>,
pub reasoning_effort: Option<String>,
+ pub speed: Option<String>,
pub provider: Option<String>,
pub provider_options: Option<serde_json::Value>,
pub metadata: Option<std::collections::HashMap<String, String>>,
@@ -318,6 +320,7 @@ impl GenerateParams {
max_tokens: None,
stop_sequences: None,
reasoning_effort: None,
+ speed: None,
provider: None,
provider_options: None,
metadata: None,
diff --git a/lib/crates/fabro-llm/src/providers/anthropic.rs b/lib/crates/fabro-llm/src/providers/anthropic.rs
index b96b8dec..b1b9dd44 100644
--- a/lib/crates/fabro-llm/src/providers/anthropic.rs
+++ b/lib/crates/fabro-llm/src/providers/anthropic.rs
@@ -109,6 +109,8 @@ struct ApiRequest {
#[serde(skip_serializing_if = "Option::is_none")]
output_config: Option<serde_json::Value>,
#[serde(skip_serializing_if = "Option::is_none")]
+ speed: Option<String>,
+ #[serde(skip_serializing_if = "Option::is_none")]
metadata: Option<std::collections::HashMap<String, String>>,
#[serde(skip_serializing_if = "std::ops::Not::not")]
stream: bool,
@@ -165,6 +167,8 @@ struct ApiUsage {
cache_read_input_tokens: Option<i64>,
#[serde(default)]
cache_creation_input_tokens: Option<i64>,
+ #[serde(default)]
+ speed: Option<String>,
}
/// Estimate reasoning tokens from thinking content blocks.
@@ -508,6 +512,7 @@ fn convert_stream_event_for_json_schema(event: StreamEvent) -> StreamEvent {
// --- Prompt caching helpers ---
const CACHE_BETA_HEADER: &str = "prompt-caching-2024-07-31";
+const FAST_MODE_BETA_HEADER: &str = "fast-mode-2026-02-01";
/// Check whether auto-caching is disabled via `provider_options`.
///
@@ -582,6 +587,7 @@ fn apply_cache_control_to_conversation_prefix(messages: &mut [ApiMessage]) {
fn build_beta_header(
provider_options: Option<&serde_json::Value>,
include_cache_header: bool,
+ include_fast_mode_header: bool,
) -> Option<String> {
let mut headers: Vec<String> = Vec::new();
@@ -604,6 +610,11 @@ fn build_beta_header(
headers.push(CACHE_BETA_HEADER.to_string());
}
+ // Add fast-mode header if speed=fast and not already present
+ if include_fast_mode_header && !headers.iter().any(|h| h == FAST_MODE_BETA_HEADER) {
+ headers.push(FAST_MODE_BETA_HEADER.to_string());
+ }
+
if headers.is_empty() {
None
} else {
@@ -698,6 +709,10 @@ impl StreamAccumulator {
self.usage.cache_write_tokens = usage
.get("cache_creation_input_tokens")
.and_then(serde_json::Value::as_i64);
+ self.usage.speed = usage
+ .get("speed")
+ .and_then(serde_json::Value::as_str)
+ .map(String::from);
}
}
vec![StreamEvent::StreamStart]
@@ -1079,6 +1094,8 @@ fn build_api_request(
.as_ref()
.map(|effort| serde_json::json!({"effort": effort}));
+ let is_fast = request.speed.as_deref() == Some("fast");
+
let api_request = ApiRequest {
model: request.model.clone(),
messages: api_messages,
@@ -1096,6 +1113,7 @@ fn build_api_request(
tool_choice: tool_choice_json,
thinking,
output_config,
+ speed: request.speed.clone(),
metadata: request.metadata.clone(),
stream,
};
@@ -1112,7 +1130,9 @@ fn build_api_request(
.header("x-api-key", &adapter.http.api_key)
.header("anthropic-version", "2023-06-01");
- if let Some(beta_str) = build_beta_header(request.provider_options.as_ref(), auto_cache) {
+ if let Some(beta_str) =
+ build_beta_header(request.provider_options.as_ref(), auto_cache, is_fast)
+ {
req_builder = req_builder.header("anthropic-beta", beta_str);
}
} else {
@@ -1196,6 +1216,7 @@ impl ProviderAdapter for Adapter {
reasoning_tokens,
cache_read_tokens: api_resp.usage.cache_read_input_tokens,
cache_write_tokens: api_resp.usage.cache_creation_input_tokens,
+ speed: api_resp.usage.speed,
..Usage::default()
},
raw: serde_json::from_str(&body).ok(),
@@ -1489,13 +1510,13 @@ mod tests {
#[test]
fn beta_header_includes_cache_header() {
- let result = build_beta_header(None, true);
+ let result = build_beta_header(None, true, false);
assert_eq!(result, Some(CACHE_BETA_HEADER.to_string()));
}
#[test]
fn beta_header_no_cache_no_user_headers() {
- let result = build_beta_header(None, false);
+ let result = build_beta_header(None, false, false);
assert_eq!(result, None);
}
@@ -1506,7 +1527,7 @@ mod tests {
"beta_headers": ["interleaved-thinking-2025-05-14"]
}
});
- let result = build_beta_header(Some(&opts), true);
+ let result = build_beta_header(Some(&opts), true, false);
assert_eq!(
result,
Some(format!(
@@ -1522,7 +1543,7 @@ mod tests {
"beta_headers": [CACHE_BETA_HEADER]
}
});
- let result = build_beta_header(Some(&opts), true);
+ let result = build_beta_header(Some(&opts), true, false);
// Should not duplicate the header
assert_eq!(result, Some(CACHE_BETA_HEADER.to_string()));
}
@@ -1534,7 +1555,7 @@ mod tests {
"beta_headers": ["interleaved-thinking-2025-05-14"]
}
});
- let result = build_beta_header(Some(&opts), false);
+ let result = build_beta_header(Some(&opts), false, false);
assert_eq!(result, Some("interleaved-thinking-2025-05-14".to_string()));
}
@@ -1586,6 +1607,7 @@ mod tests {
tool_choice: None,
thinking: None,
output_config: None,
+ speed: None,
metadata: None,
stream: false,
};
@@ -1625,6 +1647,7 @@ mod tests {
max_tokens: Some(128),
stop_sequences: None,
reasoning_effort: None,
+ speed: None,
metadata: None,
provider_options: None,
}
@@ -1941,7 +1964,7 @@ mod tests {
];
// No user headers — only cache header should appear
- let header = build_beta_header(None, true).unwrap_or_default();
+ let header = build_beta_header(None, true, false).unwrap_or_default();
for dep in &deprecated {
assert!(
!header.contains(dep),
@@ -1955,7 +1978,7 @@ mod tests {
"beta_headers": ["interleaved-thinking-2025-05-14"]
}
});
- let header = build_beta_header(Some(&opts), true).unwrap_or_default();
+ let header = build_beta_header(Some(&opts), true, false).unwrap_or_default();
for dep in &deprecated {
assert!(
!header.contains(dep),
@@ -1981,6 +2004,7 @@ mod tests {
tool_choice: None,
thinking: None,
output_config: None,
+ speed: None,
metadata: None,
stream: false,
};
@@ -2013,6 +2037,7 @@ mod tests {
tool_choice: None,
thinking: None,
output_config: None,
+ speed: None,
metadata: None,
stream: false,
};
@@ -2073,4 +2098,52 @@ mod tests {
let (api_request, _req_builder) = build_api_request(&adapter, &request, false);
assert!(api_request.output_config.is_none());
}
+
+ #[test]
+ fn build_api_request_sets_speed() {
+ let adapter = Adapter::new("test-key");
+ let request = Request {
+ speed: Some("fast".to_string()),
+ ..make_base_request()
+ };
+
+ let (api_request, _req_builder) = build_api_request(&adapter, &request, false);
+ assert_eq!(api_request.speed, Some("fast".to_string()));
+ }
+
+ #[test]
+ fn build_api_request_injects_fast_mode_beta_header() {
+ let adapter = Adapter::new("test-key");
+ let request = Request {
+ speed: Some("fast".to_string()),
+ ..make_base_request()
+ };
+
+ let (_api_request, req_builder) = build_api_request(&adapter, &request, false);
+ let built = req_builder.build().expect("should build request");
+ let beta = built
+ .headers()
+ .get("anthropic-beta")
+ .expect("anthropic-beta header should be present")
+ .to_str()
+ .unwrap();
+ assert!(
+ beta.contains(FAST_MODE_BETA_HEADER),
+ "beta header should contain fast-mode header, got: {beta}"
+ );
+ }
+
+ #[test]
+ fn beta_header_includes_both_cache_and_fast_mode() {
+ let result = build_beta_header(None, true, true);
+ let header = result.expect("should produce a header");
+ assert!(
+ header.contains(CACHE_BETA_HEADER),
+ "should contain cache header"
+ );
+ assert!(
+ header.contains(FAST_MODE_BETA_HEADER),
+ "should contain fast-mode header"
+ );
+ }
}
diff --git a/lib/crates/fabro-llm/src/providers/fabro_server.rs b/lib/crates/fabro-llm/src/providers/fabro_server.rs
index 62e66fa2..604176f8 100644
--- a/lib/crates/fabro-llm/src/providers/fabro_server.rs
+++ b/lib/crates/fabro-llm/src/providers/fabro_server.rs
@@ -245,6 +245,7 @@ mod tests {
max_tokens: None,
stop_sequences: None,
reasoning_effort: None,
+ speed: None,
metadata: None,
provider_options: None,
}
diff --git a/lib/crates/fabro-llm/src/providers/gemini.rs b/lib/crates/fabro-llm/src/providers/gemini.rs
index 306d3999..ef73894d 100644
--- a/lib/crates/fabro-llm/src/providers/gemini.rs
+++ b/lib/crates/fabro-llm/src/providers/gemini.rs
@@ -976,6 +976,7 @@ mod tests {
max_tokens: None,
stop_sequences: None,
reasoning_effort: None,
+ speed: None,
metadata: None,
provider_options: None,
}
diff --git a/lib/crates/fabro-llm/src/providers/openai.rs b/lib/crates/fabro-llm/src/providers/openai.rs
index 5d4bcccd..e13b535c 100644
--- a/lib/crates/fabro-llm/src/providers/openai.rs
+++ b/lib/crates/fabro-llm/src/providers/openai.rs
@@ -1072,6 +1072,7 @@ mod tests {
max_tokens: None,
stop_sequences: None,
reasoning_effort: None,
+ speed: None,
metadata: None,
provider_options: None,
}
diff --git a/lib/crates/fabro-llm/src/providers/openai_compatible.rs b/lib/crates/fabro-llm/src/providers/openai_compatible.rs
index 77958abe..b54ed041 100644
--- a/lib/crates/fabro-llm/src/providers/openai_compatible.rs
+++ b/lib/crates/fabro-llm/src/providers/openai_compatible.rs
@@ -1296,6 +1296,7 @@ mod tests {
max_tokens: None,
stop_sequences: None,
reasoning_effort: None,
+ speed: None,
metadata: None,
provider_options: None,
}
diff --git a/lib/crates/fabro-llm/src/types.rs b/lib/crates/fabro-llm/src/types.rs
index 05f78245..54bdf2c8 100644
--- a/lib/crates/fabro-llm/src/types.rs
+++ b/lib/crates/fabro-llm/src/types.rs
@@ -370,6 +370,8 @@ pub struct Usage {
#[serde(default, skip_serializing_if = "Option::is_none")]
pub cache_write_tokens: Option<i64>,
#[serde(default, skip_serializing_if = "Option::is_none")]
+ pub speed: Option<String>,
+ #[serde(default, skip_serializing_if = "Option::is_none")]
pub raw: Option<serde_json::Value>,
}
@@ -393,6 +395,7 @@ impl std::ops::Add for Usage {
reasoning_tokens: add_optional(self.reasoning_tokens, rhs.reasoning_tokens),
cache_read_tokens: add_optional(self.cache_read_tokens, rhs.cache_read_tokens),
cache_write_tokens: add_optional(self.cache_write_tokens, rhs.cache_write_tokens),
+ speed: self.speed,
raw: None,
}
}
@@ -452,6 +455,7 @@ pub struct Request {
pub max_tokens: Option<i64>,
pub stop_sequences: Option<Vec<String>>,
pub reasoning_effort: Option<String>,
+ pub speed: Option<String>,
pub metadata: Option<HashMap<String, String>>,
pub provider_options: Option<serde_json::Value>,
}
@@ -888,6 +892,7 @@ mod tests {
reasoning_tokens: None,
cache_read_tokens: None,
cache_write_tokens: None,
+ speed: None,
raw: None,
};
insta::assert_snapshot!(serde_json::to_string_pretty(&usage).unwrap(), @r#"
@@ -908,6 +913,7 @@ mod tests {
reasoning_tokens: Some(20),
cache_read_tokens: Some(80),
cache_write_tokens: Some(10),
+ speed: None,
raw: None,
};
insta::assert_snapshot!(serde_json::to_string_pretty(&usage).unwrap(), @r#"
@@ -940,6 +946,7 @@ mod tests {
reasoning_tokens: Some(5),
cache_read_tokens: Some(3),
cache_write_tokens: Some(1),
+ speed: None,
raw: None,
};
let b = Usage {
@@ -949,6 +956,7 @@ mod tests {
reasoning_tokens: Some(10),
cache_read_tokens: Some(7),
cache_write_tokens: Some(2),
+ speed: None,
raw: None,
};
let sum = a + b;
@@ -969,6 +977,7 @@ mod tests {
reasoning_tokens: Some(5),
cache_read_tokens: None,
cache_write_tokens: None,
+ speed: None,
raw: None,
};
let b = Usage {
@@ -978,6 +987,7 @@ mod tests {
reasoning_tokens: None,
cache_read_tokens: Some(7),
cache_write_tokens: None,
+ speed: None,
raw: None,
};
let sum = a + b;
diff --git a/lib/crates/fabro-llm/tests/integration.rs b/lib/crates/fabro-llm/tests/integration.rs
index 2aa8255d..a81118c3 100644
--- a/lib/crates/fabro-llm/tests/integration.rs
+++ b/lib/crates/fabro-llm/tests/integration.rs
@@ -15,6 +15,7 @@ fn make_request(model: &str) -> Request {
max_tokens: Some(50),
stop_sequences: None,
reasoning_effort: None,
+ speed: None,
metadata: None,
provider_options: None,
}
@@ -138,6 +139,7 @@ async fn run_multi_turn_cache_test(
max_tokens: Some(100),
stop_sequences: None,
reasoning_effort: None,
+ speed: None,
metadata: None,
provider_options: None,
};
diff --git a/lib/crates/fabro-workflows/src/backend/api.rs b/lib/crates/fabro-workflows/src/backend/api.rs
index cc751bab..d54ccb73 100644
--- a/lib/crates/fabro-workflows/src/backend/api.rs
+++ b/lib/crates/fabro-workflows/src/backend/api.rs
@@ -195,6 +195,7 @@ impl AgentApiBackend {
let config = SessionConfig {
max_tokens: node.max_tokens(),
reasoning_effort: Some(node.reasoning_effort().to_string()),
+ speed: node.speed().map(String::from),
tool_hooks,
mcp_servers,
..SessionConfig::default()
@@ -283,6 +284,7 @@ impl CodergenBackend for AgentApiBackend {
messages,
provider,
reasoning_effort: Some(node.reasoning_effort().to_string()),
+ speed: node.speed().map(String::from),
tools: None,
tool_choice: None,
response_format: None,
@@ -393,6 +395,7 @@ impl CodergenBackend for AgentApiBackend {
cache_read_tokens: response.usage.cache_read_tokens,
cache_write_tokens: response.usage.cache_write_tokens,
reasoning_tokens: response.usage.reasoning_tokens,
+ speed: response.usage.speed.clone(),
cost: None,
};
stage_usage.cost = compute_stage_cost(&stage_usage);
@@ -593,6 +596,7 @@ impl CodergenBackend for AgentApiBackend {
cache_read_tokens: total_usage.cache_read_tokens,
cache_write_tokens: total_usage.cache_write_tokens,
reasoning_tokens: total_usage.reasoning_tokens,
+ speed: total_usage.speed.clone(),
cost: None,
};
stage_usage.cost = compute_stage_cost(&stage_usage);
diff --git a/lib/crates/fabro-workflows/src/backend/cli.rs b/lib/crates/fabro-workflows/src/backend/cli.rs
index 16af8e6d..4e66884c 100644
--- a/lib/crates/fabro-workflows/src/backend/cli.rs
+++ b/lib/crates/fabro-workflows/src/backend/cli.rs
@@ -706,6 +706,7 @@ impl CodergenBackend for AgentCliBackend {
cache_read_tokens: None,
cache_write_tokens: None,
reasoning_tokens: None,
+ speed: None,
cost: None,
};
stage_usage.cost = compute_stage_cost(&stage_usage);
diff --git a/lib/crates/fabro-workflows/src/cost.rs b/lib/crates/fabro-workflows/src/cost.rs
index 09200322..0020e19a 100644
--- a/lib/crates/fabro-workflows/src/cost.rs
+++ b/lib/crates/fabro-workflows/src/cost.rs
@@ -6,9 +6,15 @@ pub fn compute_stage_cost(usage: &StageUsage) -> Option<f64> {
let info = fabro_model::get_model_info(&usage.model)?;
let input_rate = info.costs.input_cost_per_mtok?;
let output_rate = info.costs.output_cost_per_mtok?;
+ let multiplier = if usage.speed.as_deref() == Some("fast") {
+ 6.0
+ } else {
+ 1.0
+ };
Some(
- usage.input_tokens as f64 * input_rate / 1_000_000.0
- + usage.output_tokens as f64 * output_rate / 1_000_000.0,
+ (usage.input_tokens as f64 * input_rate / 1_000_000.0
+ + usage.output_tokens as f64 * output_rate / 1_000_000.0)
+ * multiplier,
)
}
@@ -47,6 +53,7 @@ mod tests {
cache_read_tokens: None,
cache_write_tokens: None,
reasoning_tokens: None,
+ speed: None,
cost: None,
};
let cost = compute_stage_cost(&usage);
@@ -63,8 +70,39 @@ mod tests {
cache_read_tokens: None,
cache_write_tokens: None,
reasoning_tokens: None,
+ speed: None,
cost: None,
};
assert_eq!(compute_stage_cost(&usage), None);
}
+
+ #[test]
+ fn compute_stage_cost_fast_mode_6x_multiplier() {
+ let standard_usage = StageUsage {
+ model: "claude-sonnet-4-5".into(),
+ input_tokens: 1000,
+ output_tokens: 500,
+ cache_read_tokens: None,
+ cache_write_tokens: None,
+ reasoning_tokens: None,
+ speed: None,
+ cost: None,
+ };
+ let fast_usage = StageUsage {
+ model: "claude-sonnet-4-5".into(),
+ input_tokens: 1000,
+ output_tokens: 500,
+ cache_read_tokens: None,
+ cache_write_tokens: None,
+ reasoning_tokens: None,
+ speed: Some("fast".into()),
+ cost: None,
+ };
+ let standard_cost = compute_stage_cost(&standard_usage).unwrap();
+ let fast_cost = compute_stage_cost(&fast_usage).unwrap();
+ assert!(
+ (fast_cost - standard_cost * 6.0).abs() < 1e-10,
+ "fast mode should be 6x standard cost: standard={standard_cost}, fast={fast_cost}"
+ );
+ }
}
diff --git a/lib/crates/fabro-workflows/src/event.rs b/lib/crates/fabro-workflows/src/event.rs
index 03c7d3bd..e8bf71f0 100644
--- a/lib/crates/fabro-workflows/src/event.rs
+++ b/lib/crates/fabro-workflows/src/event.rs
@@ -1199,6 +1199,7 @@ mod tests {
cache_read_tokens: Some(800),
cache_write_tokens: Some(50),
reasoning_tokens: Some(100),
+ speed: None,
raw: None,
},
tool_call_count: 3,
@@ -2286,6 +2287,7 @@ mod tests {
cache_read_tokens: Some(3000),
cache_write_tokens: Some(500),
reasoning_tokens: Some(800),
+ speed: None,
raw: None,
}),
};
diff --git a/lib/crates/fabro-workflows/src/outcome.rs b/lib/crates/fabro-workflows/src/outcome.rs
index 25804112..bc6313e0 100644
--- a/lib/crates/fabro-workflows/src/outcome.rs
+++ b/lib/crates/fabro-workflows/src/outcome.rs
@@ -58,6 +58,8 @@ pub struct StageUsage {
#[serde(default, skip_serializing_if = "Option::is_none")]
pub reasoning_tokens: Option<i64>,
#[serde(default, skip_serializing_if = "Option::is_none")]
+ pub speed: Option<String>,
+ #[serde(default, skip_serializing_if = "Option::is_none")]
pub cost: Option<f64>,
}
@@ -70,6 +72,7 @@ impl From<&StageUsage> for fabro_llm::types::Usage {
cache_read_tokens: u.cache_read_tokens,
cache_write_tokens: u.cache_write_tokens,
reasoning_tokens: u.reasoning_tokens,
+ speed: u.speed.clone(),
raw: None,
}
}
@@ -378,6 +381,7 @@ mod tests {
cache_read_tokens: Some(800),
cache_write_tokens: Some(50),
reasoning_tokens: Some(100),
+ speed: None,
cost: None,
};
let json = serde_json::to_string(&usage).unwrap();
@@ -398,6 +402,7 @@ mod tests {
cache_read_tokens: None,
cache_write_tokens: None,
reasoning_tokens: None,
+ speed: None,
cost: None,
};
let json = serde_json::to_string(&usage).unwrap();
diff --git a/lib/crates/fabro-workflows/src/preamble.rs b/lib/crates/fabro-workflows/src/preamble.rs
index ee8a8ab7..f000da97 100644
--- a/lib/crates/fabro-workflows/src/preamble.rs
+++ b/lib/crates/fabro-workflows/src/preamble.rs
@@ -924,6 +924,7 @@ mod tests {
cache_read_tokens: None,
cache_write_tokens: None,
reasoning_tokens: None,
+ speed: None,
cost: None,
});
outcome.files_touched = vec!["src/lib.rs".to_string(), "src/main.rs".to_string()];
@@ -1190,6 +1191,7 @@ mod tests {
cache_read_tokens: None,
cache_write_tokens: None,
reasoning_tokens: None,
+ speed: None,
cost: None,
});
node_outcomes.insert("report".to_string(), outcome);
@@ -1369,6 +1371,7 @@ mod tests {
cache_read_tokens: None,
cache_write_tokens: None,
reasoning_tokens: None,
+ speed: None,
cost: None,
});
outcome.files_touched = vec!["src/lib.rs".to_string()];
@@ -1588,6 +1591,7 @@ mod tests {
cache_read_tokens: None,
cache_write_tokens: None,
reasoning_tokens: None,
+ speed: None,
cost: None,
});
outcome.files_touched = vec!["src/lib.rs".to_string()];
diff --git a/lib/crates/fabro-workflows/src/stylesheet.rs b/lib/crates/fabro-workflows/src/stylesheet.rs
index 31fddb2d..37a3cdd8 100644
--- a/lib/crates/fabro-workflows/src/stylesheet.rs
+++ b/lib/crates/fabro-workflows/src/stylesheet.rs
@@ -2,7 +2,8 @@ use fabro_graphviz::graph::{AttrValue, Graph};
pub use fabro_graphviz::stylesheet::{parse_stylesheet, Declaration, Rule, Selector, Stylesheet};
/// Recognized stylesheet properties.
-const STYLESHEET_PROPERTIES: &[&str] = &["model", "provider", "reasoning_effort", "backend"];
+const STYLESHEET_PROPERTIES: &[&str] =
+ &["model", "provider", "reasoning_effort", "speed", "backend"];
/// Apply a stylesheet to a graph. Rules are applied by specificity order;
/// higher specificity wins. Explicit node attributes are never overridden.
@@ -240,4 +241,17 @@ mod tests {
Some(&AttrValue::String("api".into()))
);
}
+
+ #[test]
+ fn apply_speed_property() {
+ let ss = parse_stylesheet("* { speed: fast; }").unwrap();
+ let mut graph = Graph::new("test");
+ graph.nodes.insert("a".into(), Node::new("a"));
+ apply_stylesheet(&ss, &mut graph);
+
+ assert_eq!(
+ graph.nodes["a"].attrs.get("speed"),
+ Some(&AttrValue::String("fast".into()))
+ );
+ }
}
diff --git a/lib/crates/fabro-workflows/tests/integration.rs b/lib/crates/fabro-workflows/tests/integration.rs
index e1a23116..97c54fea 100644
--- a/lib/crates/fabro-workflows/tests/integration.rs
+++ b/lib/crates/fabro-workflows/tests/integration.rs
@@ -6071,6 +6071,7 @@ mod real_llm {
max_tokens: Some(200),
stop_sequences: None,
reasoning_effort: None,
+ speed: None,
metadata: None,
provider_options: None,
};

187
nodes/simplify_opus/prompt.md Normal file
View file

@ -0,0 +1,187 @@
Goal: # Plan: Support Anthropic fast mode (`speed: fast`) — Issue #120
## Context
Anthropic's API supports [fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode) (beta) for Claude Opus 4.6 — up to 2.5x faster output at 6x pricing ($30/$150 per MTok vs $5/$25 standard). The API requires:
- `"speed": "fast"` top-level field in request body
- `anthropic-beta: fast-mode-2026-02-01` header
- Response includes `usage.speed` confirming actual speed used
This follows the same pattern as `reasoning_effort`, which is already implemented end-to-end.
## Implementation
### 1. Foundation: `lib/crates/fabro-llm/src/types.rs`
- Add `speed: Option<String>` to `Request` (after `reasoning_effort`, line ~454)
- Add `speed: Option<String>` to `Usage` (before `raw`, line ~372)
- In `Usage::Add` impl (line ~389): propagate `speed` from `self` (left-hand side wins, since all requests in a session use the same speed)
### 2. Graph layer: `lib/crates/fabro-graphviz/src/graph/types.rs`
Add after `reasoning_effort()` (~line 211):
```rust
#[must_use]
pub fn speed(&self) -> Option<&str> {
self.str_attr("speed")
}
```
Returns `Option` (no default) — unlike `reasoning_effort` which defaults to "high", speed defaults to None (standard) since fast mode is 6x cost.
### 3. Stylesheet: `lib/crates/fabro-workflows/src/stylesheet.rs`
Add `"speed"` to `STYLESHEET_PROPERTIES` array (line 5):
```rust
const STYLESHEET_PROPERTIES: &[&str] = &["model", "provider", "reasoning_effort", "speed", "backend"];
```
### 4. Agent config: `lib/crates/fabro-agent/src/config.rs`
- Add `pub speed: Option<String>` field to `SessionConfig` (line ~66)
- Add `speed: None` to `Default` impl (line ~135)
### 5. Session wiring: `lib/crates/fabro-agent/src/session.rs`
- Add `speed: self.config.speed.clone()` to Request construction (~line 908)
- Add `set_speed()` method analogous to `set_reasoning_effort()` (~line 458)
### 6. Anthropic provider: `lib/crates/fabro-llm/src/providers/anthropic.rs`
**6a. `ApiRequest` struct** (~line 87): Add `speed: Option<String>` field
**6b. Beta header constant** (~line 510):
```rust
const FAST_MODE_BETA_HEADER: &str = "fast-mode-2026-02-01";
```
**6c. `build_beta_header()`** (~line 582): Add `include_fast_mode_header: bool` parameter. Inject fast-mode header (same pattern as cache header).
**6d. `build_api_request()`** (~line 1082):
- Set `speed: request.speed.clone()` on `ApiRequest`
- Compute `let is_fast = request.speed.as_deref() == Some("fast")`
- Pass `is_fast` to `build_beta_header()`
**6e. `ApiUsage`** (~line 160): Add `speed: Option<String>` field
**6f. Response parsing**: Map `api_resp.usage.speed` → `Usage.speed` in both:
- Non-streaming path (~line 1192)
- Streaming `message_start` handler (~line 690)
### 7. Usage tracking: `lib/crates/fabro-workflows/src/outcome.rs`
- Add `speed: Option<String>` to `StageUsage` (with `serde(default, skip_serializing_if)`)
- Update `From<&StageUsage> for Usage` to include `speed`
### 8. Cost multiplier: `lib/crates/fabro-workflows/src/cost.rs`
Apply 6x multiplier when `speed == "fast"`:
```rust
let multiplier = if usage.speed.as_deref() == Some("fast") { 6.0 } else { 1.0 };
// Apply multiplier to final cost
```
### 9. Backend wiring: `lib/crates/fabro-workflows/src/backend/api.rs`
- SessionConfig construction (~line 195): add `speed: node.speed().map(String::from)`
- Prompt-mode Request construction (~line 285): add `speed: node.speed().map(String::from)`
- Both StageUsage constructions (~lines 389, 589): add `speed: response.usage.speed.clone()` / `total_usage.speed.clone()`
### 10. Tests
- `anthropic.rs`: `build_api_request_sets_speed`, `build_api_request_injects_fast_mode_beta_header`, `beta_header_includes_both_cache_and_fast_mode`
- `cost.rs`: `compute_stage_cost_fast_mode_6x_multiplier`
- `stylesheet.rs`: `apply_speed_property`
- `session.rs`: `set_speed_mid_session`
- Update all existing struct literals (`Request`, `Usage`, `SessionConfig`, `StageUsage`, `ApiRequest`, `ApiUsage`) in tests to include `speed: None`
## Implementation order
1. `types.rs` (Request + Usage) — everything depends on this
2. `graph/types.rs` + `stylesheet.rs` — graph layer
3. `config.rs` + `session.rs` — agent layer
4. `anthropic.rs` — provider (ApiRequest, beta header, response parsing)
5. `outcome.rs` + `cost.rs` — usage tracking
6. `backend/api.rs` — final wiring
7. Tests throughout
## Verification
```
cargo build --workspace
cargo test --workspace
cargo fmt --check --all
cargo clippy --workspace -- -D warnings
```
## Completed stages
- **toolchain**: success
- Script: `command -v cargo >/dev/null || { curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y && sudo ln -sf $HOME/.cargo/bin/* /usr/local/bin/; }; cargo --version 2>&1`
- Stdout:
```
cargo 1.94.0 (85eff7c80 2026-01-15)
```
- Stderr: (empty)
- **preflight_compile**: success
- Script: `cargo check -q --workspace 2>&1`
- Stdout: (empty)
- Stderr: (empty)
- **preflight_lint**: success
- Script: `cargo clippy -q --workspace -- -D warnings 2>&1`
- Stdout: (empty)
- Stderr: (empty)
- **implement**: success
- Model: claude-opus-4-6, 169.9k tokens in / 37.5k out
- Files: /home/daytona/workspace/lib/crates/fabro-agent/src/compaction.rs, /home/daytona/workspace/lib/crates/fabro-agent/src/config.rs, /home/daytona/workspace/lib/crates/fabro-agent/src/session.rs, /home/daytona/workspace/lib/crates/fabro-agent/src/tools.rs, /home/daytona/workspace/lib/crates/fabro-agent/src/types.rs, /home/daytona/workspace/lib/crates/fabro-api/src/server.rs, /home/daytona/workspace/lib/crates/fabro-cli/src/doctor.rs, /home/daytona/workspace/lib/crates/fabro-graphviz/src/graph/types.rs, /home/daytona/workspace/lib/crates/fabro-hooks/src/executor.rs, /home/daytona/workspace/lib/crates/fabro-llm/src/client.rs, /home/daytona/workspace/lib/crates/fabro-llm/src/generate.rs, /home/daytona/workspace/lib/crates/fabro-llm/src/providers/anthropic.rs, /home/daytona/workspace/lib/crates/fabro-llm/src/providers/fabro_server.rs, /home/daytona/workspace/lib/crates/fabro-llm/src/providers/gemini.rs, /home/daytona/workspace/lib/crates/fabro-llm/src/providers/openai.rs, /home/daytona/workspace/lib/crates/fabro-llm/src/providers/openai_compatible.rs, /home/daytona/workspace/lib/crates/fabro-llm/src/types.rs, /home/daytona/workspace/lib/crates/fabro-llm/tests/integration.rs, /home/daytona/workspace/lib/crates/fabro-workflows/src/backend/api.rs, /home/daytona/workspace/lib/crates/fabro-workflows/src/backend/cli.rs, /home/daytona/workspace/lib/crates/fabro-workflows/src/cost.rs, /home/daytona/workspace/lib/crates/fabro-workflows/src/event.rs, /home/daytona/workspace/lib/crates/fabro-workflows/src/outcome.rs, /home/daytona/workspace/lib/crates/fabro-workflows/src/preamble.rs, /home/daytona/workspace/lib/crates/fabro-workflows/src/stylesheet.rs, /home/daytona/workspace/lib/crates/fabro-workflows/tests/integration.rs
# Simplify: Code Review and Cleanup
Review all changed files for reuse, quality, and efficiency. Fix any issues found.
## Phase 1: Identify Changes
Run git diff (or git diff HEAD if there are staged changes) to see what changed. If there are no git changes, review the most recently modified files that the user mentioned or that you edited earlier in this conversation.
## Phase 2: Launch Three Review Agents in Parallel
Use the Agent tool to launch all three agents concurrently in a single message. Pass each agent the full diff so it has the complete context.
### Agent 1: Code Reuse Review
For each change:
1. Search for existing utilities and helpers that could replace newly written code. Use Grep to find similar patterns elsewhere in the codebase — common locations are utility directories, shared modules, and files adjacent to the changed ones.
2. Flag any new function that duplicates existing functionality. Suggest the existing function to use instead.
3. Flag any inline logic that could use an existing utility — hand-rolled string manipulation, manual path handling, custom environment checks, ad-hoc type guards, and similar patterns are common candidates.
Note: This is a greenfield app, so focus on maximizing simplicity and don't worry about changing things to achieve it.
### Agent 2: Code Quality Review
Review the same changes for hacky patterns:
1. Redundant state: state that duplicates existing state, cached values that could be derived, observers/effects that could be direct calls
2. Parameter sprawl: adding new parameters to a function instead of generalizing or restructuring existing ones
3. Copy-paste with slight variation: near-duplicate code blocks that should be unified with a shared abstraction
4. Leaky abstractions: exposing internal details that should be encapsulated, or breaking existing abstraction boundaries
5. Stringly-typed code: using raw strings where constants, enums (string unions), or branded types already exist in the codebase
Note: This is a greenfield app, so be aggressive in optimizing quality.
### Agent 3: Efficiency Review
Review the same changes for efficiency:
1. Unnecessary work: redundant computations, repeated file reads, duplicate network/API calls, N+1 patterns
2. Missed concurrency: independent operations run sequentially when they could run in parallel
3. Hot-path bloat: new blocking work added to startup or per-request/per-render hot paths
4. Unnecessary existence checks: pre-checking file/resource existence before operating (TOCTOU anti-pattern) — operate directly and handle the error
5. Memory: unbounded data structures, missing cleanup, event listener leaks
6. Overly broad operations: reading entire files when only a portion is needed, loading all items when filtering for one
## Phase 3: Fix Issues
Wait for all three agents to complete. Aggregate their findings and fix each issue directly. If a finding is a false positive or not worth addressing, note it and move on — do not argue with the finding, just skip it.
When done, briefly summarize what was fixed (or confirm the code was already clean).

5
nodes/simplify_opus/provider_used.json Normal file
View file

@ -0,0 +1,5 @@
{
"mode": "agent",
"provider": "anthropic",
"model": "claude-opus-4-6"
}

27
nodes/simplify_opus/response.md Normal file
View file

@ -0,0 +1,27 @@
All checks pass. Here's my summary of findings from the three parallel reviews:
## Summary
**The code is clean — no issues to fix.**
### Agent 1: Code Reuse Review
- **No duplication found.** The `speed` field follows the exact same pattern as `reasoning_effort` throughout the codebase (structs, config, session, graph node accessor, stylesheet, backend wiring). This is consistent, not duplicative.
- **`as_deref() == Some("fast")` appears in 2 places** (anthropic.rs and cost.rs) — but they serve different purposes in different crates. Extracting a shared helper would be over-engineering.
- **`set_speed()` follows `set_reasoning_effort()` pattern** — consistent.
### Agent 2: Code Quality Review
- **`build_beta_header` has 3 params (2 booleans)** — borderline parameter sprawl, but it's a private function with a single callsite. The two booleans are independent concerns (caching vs fast-mode) that don't warrant a struct.
- **`speed: Option<String>` not an enum** — matches the existing `reasoning_effort: Option<String>` convention and the API's string representation. Consistent.
- **`speed` on `Usage` struct** — confirmed correct per Anthropic API docs: the response `usage` object includes `"speed": "fast"`. This is where the API returns it.
- **`Usage::Add` takes LHS speed** — correct, since all requests in a session use the same speed setting.
### Agent 3: Efficiency Review
- **String clones** (`Option<String>` with value "fast") are trivially cheap per-request — not a concern.
- **Linear scan in `build_beta_header`** on a Vec with at most ~3 elements — negligible.
- **No hot-path bloat, no N+1 patterns, no unnecessary allocations.**
### Verification
- `cargo build --workspace` ✅
- `cargo test --workspace` ✅ (179 passed, 0 failed)
- `cargo clippy --workspace -- -D warnings` ✅
- `cargo fmt --check --all` ✅

6
nodes/simplify_opus/status.json Normal file
View file

@ -0,0 +1,6 @@
{
"status": "success",
"notes": "Stage completed: simplify_opus",
"failure_reason": null,
"timestamp": "2026-03-20T22:28:21.029938+00:00"
}