ai-dispatch 8.68.0

Multi-AI CLI team orchestrator
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157

// Batch task -> run args conversion helpers.
// Exports: task_to_run_args
// Deps: crate::cmd::run::RunArgs, crate::batch, crate::store::Store
use crate::batch;
use crate::cmd::run::{RunArgs, NO_SKILL_SENTINEL};
use crate::store::Store;
use std::collections::HashMap;
use std::sync::Arc;

pub(crate) fn task_to_run_args(
    task: &batch::BatchTask,
    siblings: &[&batch::BatchTask],
    background: bool,
    store: &Arc<Store>,
    shared_dir_path: Option<&str>,
) -> RunArgs {
    // If team is set and agent is empty/auto, auto-select from team members
    let agent_name = if (task.agent.is_empty() || task.agent == "auto") && task.team.is_some() {
        let team_config = task.team.as_deref().and_then(crate::team::resolve_team);
        let selection_opts = crate::agent::RunOpts {
            dir: task.dir.clone(),
            output: task.output.clone(),
            model: task.model.clone(),
            budget: task.budget,
            read_only: task.read_only,
            context_files: vec![],
            session_id: None,
            env: None,
            env_forward: None,
        };
        let (selected, reason) = crate::agent::select_agent_with_reason(
            &task.prompt,
            &selection_opts,
            store,
            team_config.as_ref(),
        );
        aid_info!("[aid] Batch auto-selected: {selected} (reason: {reason})");
        selected
    } else if task.agent.is_empty() {
        "auto".to_string()
    } else {
        task.agent.clone()
    };
    let batch_siblings = siblings
        .iter()
        .map(|sibling| {
            (
                sibling
                    .name
                    .clone()
                    .or_else(|| sibling.id.clone())
                    .unwrap_or_else(|| "<unnamed>".to_string()),
                if sibling.agent.is_empty() {
                    "auto".to_string()
                } else {
                    sibling.agent.clone()
                },
                sibling.prompt.clone(),
            )
        })
        .collect();
    let cascade = task
        .fallback
        .as_deref()
        .map(|fallback| {
            fallback
                .split(',')
                .map(str::trim)
                .filter(|agent| !agent.is_empty())
                .map(ToString::to_string)
                .collect()
        })
        .unwrap_or_else(|| auto_cascade_for_rate_limited(&agent_name));
    let env = crate::idle_timeout::env_with_idle_timeout(
        merged_env(task.env.as_ref(), task.env_forward.as_ref(), shared_dir_path),
        task.idle_timeout,
    );
    let skills = if task.no_skill {
        vec![NO_SKILL_SENTINEL.to_string()]
    } else {
        task.skills.clone().unwrap_or_default()
    };
    RunArgs {
        agent_name,
        prompt: task.prompt.clone(),
        dir: task.dir.clone(),
        output: task.output.clone(),
        result_file: task.result_file.clone(),
        model: task.model.clone(),
        worktree: task.worktree.clone(),
        group: task.group.clone(),
        container: task.container.clone(),
        verify: task.verify.clone(),
        iterate: task.iterate,
        eval: task.eval.clone(),
        eval_feedback_template: task.eval_feedback_template.clone(),
        judge: task.judge.clone(),
        peer_review: task.peer_review.clone(),
        max_duration_mins: task.max_duration_mins.map(|value| value as i64),
        retry: task.retry.unwrap_or(0),
        context: task.context.clone().unwrap_or_default(),
        checklist: task.checklist.clone().unwrap_or_default(),
        skills,
        hooks: task.hooks.clone().unwrap_or_default(),
        background,
        dry_run: false,
        announce: true,
        on_done: task.on_done.clone(),
        cascade,
        read_only: task.read_only,
        sandbox: task.sandbox,
        budget: task.budget,
        best_of: task.best_of,
        metric: task.metric.clone(),
        team: task.team.clone(),
        context_from: task.context_from.clone().unwrap_or_default(),
        batch_siblings,
        scope: task.scope.clone().unwrap_or_default(),
        parent_task_id: task.parent.clone(),
        existing_task_id: task.id.as_ref().map(|id| crate::types::TaskId(id.clone())),
        env,
        env_forward: task.env_forward.clone(),
        ..Default::default()
    }
}

/// If the agent is rate-limited, return the suggested fallback as an auto-cascade.
fn auto_cascade_for_rate_limited(agent_name: &str) -> Vec<String> {
    let Some(agent) = crate::types::AgentKind::parse_str(agent_name) else {
        return vec![];
    };
    if !crate::rate_limit::is_rate_limited(&agent) {
        return vec![];
    }
    crate::agent::selection::coding_fallback_for(&agent)
        .map(|fallback| vec![fallback.as_str().to_string()])
        .unwrap_or_default()
}

fn merged_env(
    env: Option<&HashMap<String, String>>,
    env_forward: Option<&Vec<String>>,
    shared_dir_path: Option<&str>,
) -> Option<HashMap<String, String>> {
    let mut merged = env.cloned().unwrap_or_default();
    if let Some(shared_dir_path) = shared_dir_path {
        merged.insert("AID_SHARED_DIR".to_string(), shared_dir_path.to_string());
    }
    if let Some(env_forward) = env_forward {
        for name in env_forward {
            if let Ok(value) = std::env::var(name) {
                merged.insert(name.clone(), value);
            }
        }
    }
    (!merged.is_empty()).then_some(merged)
}