| name | auto-remediation-ecosystem |
| version | 1.0.0 |
| author | Hermes Cortex |
| created | 2026-06-15T00:00:00.000Z |
| updated | 2026-06-15T00:00:00.000Z |
| tags | ["auto-remediation","ecosystem","setup","troubleshooting","monitoring","server-administration","automation","production-operations"] |
| description | Complete auto-remediation ecosystem setup, configuration, and maintenance
for production systems. Covers deployment of all auto-remediation components,
integration with existing server administration workflows, monitoring,
troubleshooting, and production operations best practices.
|
| syntax | # Install the ecosystem setup
hermes skills install auto-remediation-ecosystem
# Verify complete system
ecosystem-verify.sh
# Setup reference
cat .hermes/skills/auto-remediation-ecosystem/references/auto-remediation-ecosystem.md
|
| references | ["references/auto-remediation-ecosystem.md","templates/auto-remediation-full-setup.sh","scripts/ecosystem-verify.sh","scripts/auto-remediation-health-check.sh","scripts/manual-triggers-guide.sh"] |
| examples | [{"title":"Verify a Fresh Ecosystem Install","content":"```bash\necosystem-verify.sh\n# Expect: all components present, services healthy, cron registered\n```\n"},{"title":"Run the Health Check","content":"```bash\nauto-remediation-health-check.sh\n# Reports per-component status: sensor, fixer, queue, delivery\n```\n"}] |
Auto-Remediation Ecosystem
Overview
The auto-remediation ecosystem is the full production pipeline that detects
operational failures, remediates them automatically, and reports the outcome.
This skill covers deployment, configuration, monitoring, and maintenance
of the complete ecosystem — not just a single component.
Components
| Component | Role | Entry point |
|---|
| Sensor | Detects failures (cron failures, service down, inbox drift) | agent-remediation-sensor cron |
| Fixer | Applies remediations autonomously | agent-fixer-* crons |
| Queue / state | Tracks what failed and what was attempted | PGMQ / state files |
| Delivery | Reports outcomes to the orchestrator / user | Agent bus + Telegram |
| Verification | Confirms the fix worked | re-run the failing check |
Deployment
bash templates/auto-remediation-full-setup.sh
bash scripts/ecosystem-verify.sh
cronjob action=list | grep remediation
Deployment prerequisites (per component):
- Passwordless sudo for the fixer to act (documented in the setup skill)
- Agent bus access for delivery to the orchestrator
- Cron scheduler healthy — the doctor validates expected crons
Monitoring
Run the health check on a schedule or on demand:
bash scripts/auto-remediation-health-check.sh
Watch for:
- Sensor silence — a sensor that stops firing means it stopped running,
not that everything is fine (check the cron's
last_status).
- Fixer loops — the same failure remediated repeatedly without resolution
is a failed fix, not a healthy loop. Escalate after N retries.
