| name | status |
| description | Read a Grok run by id: finished envelope, or running progress if still in progress (read-only) |
| argument-hint | [--run-id <run-id> | <run-id>] |
| allowed-tools | Bash(node:*) |
How to run (transparent)
- Take the absolute Base directory for this skill from the Skill tool
(the folder that contains this skill's
SKILL.md and run.mjs).
- Set
SKILL_BASE to that path. Do not invent versioned cache paths.
- Always invoke the companion only through this skill's runner:
SKILL_BASE='<Base directory for this skill - absolute path from Skill tool>'
node "$SKILL_BASE/run.mjs" <mode> [args...]
run.mjs finds the plugin install from its own location and runs
scripts/grok-companion.mjs. No CLAUDE_PLUGIN_ROOT / PLUGIN_ROOT required.
If the host already exported CLAUDE_PLUGIN_ROOT or PLUGIN_ROOT, you may call
node "$CLAUDE_PLUGIN_ROOT/scripts/grok-companion.mjs" instead; prefer
"$SKILL_BASE/run.mjs" whenever the Skill tool loaded this skill.
Return companion stdout verbatim. Never put free-text in --task "...";
use --task-file - with a single-quoted heredoc.
Read a Grok run by id through the hardened wrapper and relay its result
envelope. Projection (read-only; never writes the target run dir):
- In-flight (
created / running / finalizing) → top-level "running", exit 0
- Completed →
"success", exit 0, with response.storedEnvelope when present
- Failed / canceled / derived interrupted →
"failure", exit 1, still a
well-formed status envelope (exit 1 means the target failed or was
interrupted, not that the status command malfunctioned)
response.target includes lifecycle, lifecycleSource (record |
envelope | derived), elapsedMs, process liveness, and progress summaries.
Always relay the JSON envelope to the user regardless of exit status.
Raw slash-command arguments:
$ARGUMENTS
Required wrapper flags (copy exactly, substitute only placeholder values):
--run-id <run-id> is required. If the user did not supply one, ask them for
the run id (it is printed in every run's envelope as runId) BEFORE running
anything.- Preserve the user's arguments exactly. Do not strip, add, or reorder flags.
Do not invent a flag that is not in the argument-hint.
Run it as one Bash call and relay the result.
- Injection safety (canonical rationale:
plugin/references/argv-safety.md):
Wrap every substituted flag VALUE in single quotes (--run-id '<run-id>').
Bare flags (--web) carry no value to quote. The wrapper then rejects any run
id that is not the strict YYYYMMDDThhmmssZ-xxxxxx run-id shape:
node "$SKILL_BASE/run.mjs" status --run-id '<run-id from $ARGUMENTS>'
- Return the command stdout envelope to the user VERBATIM. Do not paraphrase,
summarize, reformat, or add commentary before or after it. Preserve the exit
status: exit 1 with a parseable status envelope is a successful inspection of
a failed/interrupted target - still show the envelope.
- Harness distinction: if top-level
error is set (invalid-target,
output-malformed, …), the status command could not inspect the run. If
response.target.lifecycle is present with failed/canceled/interrupted and
there is no error field, inspection succeeded and the target is bad.
If the companion prints an actionable "could not locate the Grok wrapper"
message instead of an envelope, tell the user to run /grok:setup.
Without --run-id, the companion prints the local jobs table for this workspace
(recent companion-tracked runs). With --run-id, it returns the wrapper status envelope.
A bare positional matching the runId shape is accepted as well - the companion
translates it to --run-id <id> before the wrapper runs.
Also: /grok:jobs, /grok:result, /grok:cancel.
