0 on success,
non-zero on error.
archal trace
Fetch scenario traces from the Archal control plane (api.archal.ai). This is
a remote lookup — requires archal login or ARCHAL_TOKEN — and returns only
traces you own. It is not a local-cache reader; local artifacts live under
.archal/cache/ and are inspected with cat/jq, not this command.
| Flag | Description | Default |
|---|---|---|
-n, --limit <count> | Number of traces to list | 20 |
archal trace <name> to see the
detail view, which adds scenario title, error count, and a numbered event
list of span tool calls (relative timestamp, twin.tool, and a short
output-or-error summary capped at 80 chars). A 401/403 from the API means
your token is missing or scoped to a different user.
archal inspect session <id>
Print a hosted session as JSON: resolved services, seeds, TTL, twin URLs, and
worker status. Useful when debugging a CI failure against a specific session
id from the dashboard.
archal inspect run <id>
Print a managed run bundle as JSON — the run’s scenario, criteria, per-run
scores, and trace metadata.
archal export run <id>
Export the full artifact bundle for a run, including every tool call and state
snapshot, as a single JSON document. Pipe to a file when you want a portable
record or want to share a failing run.
| Flag | Description |
|---|---|
--anonymize | Redact sensitive evidence payload fields before writing |
archal usage
Print your account’s session-minutes usage for the current billing period and
current plan.
archal doctor
Probe the active twin session end-to-end: reachability, auth, URL resolution,
and tool surface per twin. Use this first when “nothing works”.
| Flag | Description |
|---|---|
--twin <name> | Probe a single twin by id (defaults to all twins in the active session) |
--json | Emit diagnosis as JSON |