From oh-my-evor
Load and validate the mission contract, set active run state, and invoke the Evor tick loop
How this skill is triggered — by the user, by Claude, or both
Slash command
/oh-my-evor:evor-run [mission-id or run-id][mission-id or run-id]The summary Claude sees in its skill listing — used to decide when to auto-load this skill
<Purpose>
<Use_When>
/evor-runevor-setup completes and prints "Start the tick loop with /evor-run"<Do_Not_Use_When>
evor-setup insteadevor-resume which accepts a run-idevor-dashboard
</Do_Not_Use_When>If arguments were provided, treat them as mission-id or run-id. Otherwise:
evor_state_read to check for a current active run.status != "completed": offer to resume that run.evor_state_read and prompt the user to select one, or redirect to /evor-setup.Call evor_read_goal_contract({ run_id: "<run_dir>" }) to load and validate the contract.
If it reports missing or invalid fields, print the error and redirect to /evor-setup. Do not enumerate specific field names — let the tool report what is wrong.
Call evor_state_read to check whether the mission is locked.
ERROR: Mission state not found — mission is not locked.
The contract has not passed the Phase-2 validation gate.
Run /evor-validate to validate and lock it, or /evor-setup to reinitialize.
ERROR: Mission is not locked.
The contract has not passed the Phase-2 validation gate.
Run /evor-validate to validate and lock it, or /evor-setup to reinitialize.
Call evor_state_read to check run state:
Resuming mission: <mission_id>
Run ID: <run_id>
Ticks completed: <tick_count>
Best score so far: <best_score> (baseline: <baseline_value>)
Frontier size: <len(frontier_ids)>
Last eval version: <current_eval_version>
Ask: "Resume from tick <tick_count + 1>? (yes/no)"
→ "no": allow the user to specify a different run or abort.status = "completed": print "This run is already complete. Use /evor-report to view results or /evor-setup to start a new mission." and stop.Call evor_state_write to record the active run and set status to running:
evor_state_write({
mission_status: "running",
active_run: {
mission_id: "<mission_id>",
run_id: "<run_id>",
run_dir: "<run_dir>",
started_at: "<ISO 8601>",
status: "running"
}
})
Choose the appropriate run-watch mode based on context:
evor_run_start fires inside the tick loop, the evor skill watches the job with the native Monitor tool (tail -f <log_path> | grep -E --line-buffered "elapsed_steps=|val_|Traceback|Error|OOM|Killed|FAILED"). Session stays live.watchPaths at run_start) fires a FileChanged wake when jobs/<id>/status.json changes. Use CronCreate to establish a recurring check-in cadence for multi-day missions. Session sleeps between check-ins.The tick loop in the evor skill manages this automatically; this note is for orientation when choosing /evor-run vs a scheduled invocation.
Read and follow skills/evor/SKILL.md exactly, passing:
run_id = the resolved run IDresume = true if tick_count > 0The evor skill owns the tick loop from this point forward.
<Tool_Usage>
evor_state_read — read active run state, mission lock status, and run progressevor_read_goal_contract — load and validate the mission contractevor_state_write — set mission_status="running" and active_run recordevor — hand off tick loop execution
</Tool_Usage>npx claudepluginhub it-dainb/oh-my-evorGuides collaborative design exploration before implementation: explores context, asks clarifying questions, proposes approaches, and writes a design doc for user approval.
Creates structured, bite-sized implementation plans from specs or requirements before writing code. Useful for breaking down multi-step tasks into testable steps with file structure and task boundaries.
Synthesizes the current conversation into a structured spec (PRD) and publishes it to the project issue tracker with a ready-for-agent label, without interviewing the user.