Make the brief
explicit.

Run flags

-q, --quiet

Suppress informational output (only errors/warnings shown). Also set via AID_QUIET=1

--repo <REPO>

Or --repo-root supplies the repository anchor

-d, --dir <DIR>

Sets the task working directory

-o, --output <OUTPUT>

Selects a task output path

--result-file <FILE>

Requires a durable result artifact

-m, --model <MODEL>

aid validates requested --model parameters against the target CLI's served model list

--difficulty <DIFFICULTY>

[possible values: trivial, simple, moderate, complex]

--budget <BUDGET>

[possible values: free, cheap, standard, premium]

--urgency <URGENCY>

[possible values: background, normal, urgent]

--rigor <RIGOR>

Proof level; critical requires enabled --verify and post-task --audit (or project defaults) [possible values: draft, standard, critical]

--egress <EGRESS>

Data egress: any (default), local (loopback only), or private-network (loopback or RFC1918/link-local) Possible values: - any: Any provider (default). No data-locality gate - local: Only a provider whose established endpoint is loopback - private-network: Only a provider whose established endpoint is loopback or private-network [default: any]

--kind <KIND>

Task category; for a bug audit use debugging --read-only (--audit is a post-task check) [possible values: research, simple-edit, complex-impl, frontend, debugging, testing, refactoring, documentation]

-w, --worktree <WORKTREE>

Create/reuse a writable task branch; for a read-only checkout use --dir <path>

-g, --group <GROUP>

Assign to workgroup

--verify [<VERIFY>]

Eligible plain cargo check|test|clippy without shell syntax uses aid's build runner outside remote-build/container runs, adding --message-format=json; everything else uses sh -c in the task directory; bare --verify auto-detects, an absent flag uses the project default, and a program missing under sh -c fails verification (missing cargo on the runner is infrastructure)

--judge [<AGENT>]

AI judge evaluates output quality

--peer-review <AGENT>

Dispatch peer review after completion

--retry <RETRY>

Permits new attempts after failure

--context <CONTEXT>...

for source material, not extra positional arguments

--checklist <CHECKLIST>...

or --checklist-file for explicit acceptance criteria

--scope <SCOPE>...

to state intended files

--remote-build [<BOX>]

Run Cargo builds on an rbox build box (bare flag selects automatically)

--context-from <CONTEXT_FROM>...

Inject output from previous task(s) as context

--skill <SKILL>...

Methodology skills to inject

--template <TEMPLATE>

Prompt template to wrap around the task

--on-done <ON_DONE>

Command to run on task completion

--cascade <CASCADE>

Agent cascade: comma-separated list of agents to try on failure (e.g. opencode,codex,cursor)

--hook <HOOK>

Hook specs to run for the dispatched task

--backup <TARGET[:FOLDER]>

Back up task artifacts when the task ends: TARGET[:FOLDER], e.g. gdrive:audits/{project}

--no-backup

Skip artifact backup for this task even when the project configures one

--bg

Returns the task ID immediately

--read-only

Read-only task; use --dir <checkout-path> to inspect an existing worktree

--sandbox

Requests sandboxed execution

--container <IMAGE>

Run agent inside a container

--best-of <N>

races the top N launchable builtin aid advise candidates for the declared profile, kind, team and caller pool

--metric <COMMAND>

Custom metric command for best-of selection

--timeout <SECS>

Hard cap for the run in seconds; for minutes, multiply by 60

--idle-timeout <SECS>

Kill the agent after this many seconds without parsed activity

--audit

Run aic cross-audit on this task after completion (requires `aic` binary on PATH)

--no-audit

Skip aic cross-audit for this task, even if enabled by project defaults