automation-run

Scheduled Automation Runner

Use the current schedule and the installed automation package's owner documentation to identify and run the requested jobs. This entry does not hardcode package, repository, or command names.

Execution workflow

Phase 1: Identify the requested jobs

• Resolve the current repository from the invocation or gh repo view.
• Inspect the schedule without printing raw crontab lines or credential values.
• Report only the schedule, job identifier, executable, and non-secret arguments; preserve all existing entries, including GITHUB_TOKEN lines.
• Use the scheduler's current configuration as the source of truth; do not use historical examples as a job inventory.

Phase 2: Run the selected jobs

• Follow the installed package's supported invocation for each selected job.
• Keep credentials in the existing process environment and out of command arguments, logs, and terminal output.
• Capture the actual command result and exit status. Do not kill, restart, or duplicate unrelated scheduled processes.

Phase 3: Verify the requested result

• Check the selected job's concrete output or artifact using its owning workflow.
• A process start, exit code alone, or generic success message does not prove the requested result.
• Report the verified result and any unresolved part as UNVERIFIED.