Eve Job Debugging

When your Eve job is stuck, failing, or you just need to know what's happening, this skill walks you through the CLI debugging commands that replicate the client experience. It covers `eve job follow`, `eve job diagnose`, and `eve job wait` for monitoring, plus the less obvious stuff like per-job harness overrides that can silently change behavior at create time. The documentation also explains the orchestrator's self-healing for pod loss and why some jobs that used to get stuck no longer do. If you're reaching for kubectl to debug an Eve job, this skill reminds you that the answer should be in the CLI, and shows you where to look.

skills.sh installs: 312
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add incept5/eve-skillpacks --skill eve-job-debugging --agent claude-code

Installs into .claude/skills of the current project.

Files

Eve Job Debugging

CLI-Only Debugging

Debug via the Eve CLI exclusively. This replicates the client experience — clients don't have kubectl or host access.

Every debugging capability must be available through the CLI. If you find yourself needing system tools to diagnose a job issue, that's a gap in our CLI that should be fixed.

Monitor

eve job follow <id> to stream logs.
eve job wait <id> --timeout 300 --json to wait on completion.
eve job result <id> --format text for the latest result.

Diagnose

eve job diagnose <id> for timeline and error summary.
eve job show <id> --verbose for attempts and phase.
eve job dep list <id> for dependency blocks.

Per-Job Harness Overrides

When debugging unexpected harness behavior, suspect a per-job override before suspecting agent defaults. Jobs may carry inline overrides set at create time:

--harness-override-file <path.json> — inline {harness, model?, reasoning_effort?, variant?, temperature?} bundle that wins over harness_profile.
--env-override KEY=VALUE (repeatable) — per-job env values, may include ${secret.KEY} placeholders resolved at spawn.

SKILL.md on GitHub