Automated postmortem scanner for detecting and reporting recurring cron job failures.

Install

mkdir -p .claude/skills/drift-watcher && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/11358" && unzip -o skill.zip -d .claude/skills/drift-watcher && rm skill.zip

Installs to .claude/skills/drift-watcher

Activation

This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.

Hourly cron-postmortem sweep — scans launchd job logs for failure signatures and emits Drift incidents for routing through the incident pipeline.
145 charsno explicit “when” trigger
Intermediate

Key capabilities

  • Scan launchd job logs
  • Detect cron failure signatures
  • Emit Drift incident records
  • Reset status on job recovery

How it works

The skill runs an hourly sweep of launchd logs, matches output against known failure signatures, and emits incidents to the Drift pipeline.

Inputs & outputs

You give it
launchd stdout/stderr logs
You get back
Structured Drift incident records

When to use drift-watcher

  • Monitoring cron job health
  • Reporting recurring failures
  • System-level task observability

About this skill

drift-watcher — hourly cron health sweep

Phase 1 of the Drift incident ledger: detect cron failures by reading launchd log output and emit structured incident records.

When this runs

  • Primary: cron. ai.openclaw.drift-watcher.plist fires hourly 09:00–00:00 local.
  • Wrapper: through skill-wrapper --skill drift-watcher --agent main --trigger cron.

What it does

Invokes ${OPENCLAW_HOME}/bin/cron-postmortem-watch:

  1. Read recent launchd stdout/stderr logs for known plists.
  2. Match against failure signatures (Nth consecutive error, timeout, missing artifact, etc.).
  3. For each match, call incident emit with the appropriate taxonomy.
  4. Apply green-path counter reset when a previously-failing job recovers.

Exit codes

  • 0 — sweep completed (incidents emitted as needed)
  • non-zero — script-level error reading logs

Files

  • Script: ${OPENCLAW_HOME}/bin/cron-postmortem-watch
  • Plist: $HOME/Library/LaunchAgents/ai.openclaw.drift-watcher.plist
  • This spec: ${OPENCLAW_HOME}/skills/drift-watcher/SKILL.md

Related

  • incident-emit — direct downstream.
  • incident-analyze / incident-notify — phases 2/3.
  • drift-learn — phase 4 (weekly aggregation).

When not to use it

  • When monitoring non-launchd tasks
  • When the Drift incident pipeline is unavailable

Prerequisites

launchd job logsOPENCLAW_HOME environment variable

Limitations

  • Limited to hourly execution schedule
  • Requires access to specific plist files

How it compares

It automates the postmortem detection process for cron jobs rather than relying on manual log inspection.

Compared to similar skills

drift-watcher side by side with the closest alternatives in the catalog.

SkillInstallsUpdatedSafetyDifficulty
drift-watcher (this skill)01moNo flagsIntermediate
system-info16moReviewBeginner
it-operations17moNo flagsAdvanced
pagerduty-automation05moNo flagsIntermediate

Try saying

Example prompts that trigger this skill in your AI assistant.

Search skills

Search the agent skills registry