Struct AgentConfig 

pub struct AgentConfig {

    pub runner: Option<Runner>,
    pub model: Option<Model>,
    pub reasoning_effort: Option<ReasoningEffort>,
    pub iterations: Option<u8>,
    pub followup_reasoning_effort: Option<ReasoningEffort>,
    pub codex_bin: Option<String>,
    pub opencode_bin: Option<String>,
    pub gemini_bin: Option<String>,
    pub claude_bin: Option<String>,
    pub cursor_bin: Option<String>,
    pub kimi_bin: Option<String>,
    pub pi_bin: Option<String>,
    pub claude_permission_mode: Option<ClaudePermissionMode>,
    pub runner_cli: Option<RunnerCliConfigRoot>,
    pub phase_overrides: Option<PhaseOverrides>,
    pub instruction_files: Option<Vec<PathBuf>>,
    pub repoprompt_plan_required: Option<bool>,
    pub repoprompt_tool_injection: Option<bool>,
    pub ci_gate: Option<CiGateConfig>,
    pub git_revert_mode: Option<GitRevertMode>,
    pub git_commit_push_enabled: Option<bool>,
    pub phases: Option<u8>,
    pub notification: NotificationConfig,
    pub webhook: WebhookConfig,
    pub session_timeout_hours: Option<u64>,
    pub scan_prompt_version: Option<ScanPromptVersion>,
    pub runner_retry: RunnerRetryConfig,
}

Agent runner defaults (Claude, Codex, OpenCode, Gemini, or Cursor).

Fields

runner: Option<Runner>

Which harness to use by default.

model: Option<Model>

Default model.

reasoning_effort: Option<ReasoningEffort>

Default reasoning effort (only meaningful for Codex models).

iterations: Option<u8>

Number of iterations to run for each task (default: 1).

followup_reasoning_effort: Option<ReasoningEffort>

Reasoning effort override for follow-up iterations (iterations > 1). Only meaningful for Codex models.

codex_bin: Option<String>

Override the codex executable name/path (default is “codex” if None).

opencode_bin: Option<String>

Override the opencode executable name/path (default is “opencode” if None).

gemini_bin: Option<String>

Override the gemini executable name/path (default is “gemini” if None).

claude_bin: Option<String>

Override the claude executable name/path (default is “claude” if None).

cursor_bin: Option<String>

Override the cursor agent executable name/path (default is “agent” if None).

NOTE: Cursor’s runner binary name is agent (not cursor).

kimi_bin: Option<String>

Override the kimi executable name/path (default is “kimi” if None).

pi_bin: Option<String>

Override the pi executable name/path (default is “pi” if None).

claude_permission_mode: Option<ClaudePermissionMode>

Claude permission mode for tool and edit approval. AcceptEdits: auto-approves file edits only BypassPermissions: skip all permission prompts (YOLO mode)

runner_cli: Option<RunnerCliConfigRoot>

Normalized runner CLI behavior overrides (output/approval/sandbox/etc).

This is additive: existing runner-specific fields remain supported.

phase_overrides: Option<PhaseOverrides>

Per-phase overrides for runner, model, and reasoning effort.

Allows specifying different settings for each phase (1, 2, 3). Phase-specific values override the global agent settings.

instruction_files: Option<Vec<PathBuf>>

Additional instruction files to inject at the top of every prompt sent to runner CLIs.

Paths may be absolute, ~/-prefixed, or repo-root relative. Missing files are treated as configuration errors. To include repo-local AGENTS.md, add "AGENTS.md" to this list.

repoprompt_plan_required: Option<bool>

Require RepoPrompt usage during planning (inject context_builder instructions).

repoprompt_tool_injection: Option<bool>

Inject RepoPrompt tooling reminder block into prompts.

ci_gate: Option<CiGateConfig>

Structured CI gate execution settings.

git_revert_mode: Option<GitRevertMode>

Controls automatic git revert behavior when runner or supervision errors occur.

git_commit_push_enabled: Option<bool>

Enable automatic git commit and push after successful runs (default: true).

phases: Option<u8>

Number of execution phases (1, 2, or 3). 1 = single-pass, 2 = plan+implement, 3 = plan+implement+review.

notification: NotificationConfig

Desktop notification configuration for task completion.

webhook: WebhookConfig

Webhook configuration for HTTP task event notifications.

session_timeout_hours: Option<u64>

Session timeout in hours for crash recovery (default: 24). Sessions older than this threshold are considered stale and require explicit user confirmation to resume.

scan_prompt_version: Option<ScanPromptVersion>

Scan prompt version to use (v1 or v2, default: v2).

runner_retry: RunnerRetryConfig

Runner invocation retry/backoff configuration.

Implementations

impl AgentConfig

pub fn ci_gate_enabled(&self) -> bool

pub fn ci_gate_display_string(&self) -> String

pub fn merge_from(&mut self, other: Self)

Trait Implementations

impl Clone for AgentConfig

fn clone(&self) -> AgentConfig

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for AgentConfig

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for AgentConfig

fn default() -> AgentConfig

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for AgentConfig

where AgentConfig: Default,

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl JsonSchema for AgentConfig

fn schema_name() -> Cow<'static, str>

The name of the generated JSON Schema.

fn schema_id() -> Cow<'static, str>

Returns a string that uniquely identifies the schema produced by this type.

fn json_schema(generator: &mut SchemaGenerator) -> Schema

Generates a JSON Schema for this type.

fn inline_schema() -> bool

Whether JSON Schemas generated for this type should be included directly in parent schemas, rather than being re-used where possible using the $ref keyword.

impl Serialize for AgentConfig

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.