metallurgical /ogre:task-list
Use this skill when the user wants to see all steps/tasks that belong to one job, not just a single task or the whole-runtime status.
Inputs
/ogre:task-list <job-id>— e.g./ogre:task-list job-abc123
Job id required. If the user gives an issue slug/number instead, run /ogre:status <issue> first to read its Job Id row, then call this with that job id.
Behavior
Run:
${CLAUDE_PLUGIN_ROOT}/scripts/ogre task-list <job-id>
This also re-syncs the plan first, so every checklist step shows up here — including ones nobody has run execute against yet (those show pending, Executor: -).
Output
Job Id: job-<uuid> Issue: <slug>
+---+----------+---------+-----------------+--------------------------+
| # | Task Id | Status | Executor | Step |
+---+----------+---------+-----------------+--------------------------+
| 1 | task-... | passed | codex (gpt-5.5) | Step one: create file A… |
| 2 | task-... | pending | - | Step two: create file B |
+---+----------+---------+-----------------+--------------------------+
View one: ogre status --task <task-id>
Sorted by # (step_index, the checklist's own order) — step 1 first, not newest-first. One row per checklist item in the plan, always, whether or not it's been executed. Status one of pending/running/passed/failed/stopped. Long step text shortened to ~60 chars with a trailing …. If the job has no checklist synced yet, prints No tasks yet for this job. Errors out if the job id doesn't resolve to any issue.
Show the table verbatim, in a code block, exactly as the helper printed it — same rule as /ogre:status: do not paraphrase rows into prose, do not drop columns. Add commentary after the block, not instead of it.