JPeetz/agent-skills / skills/devops-infrastructure/maintenance
maintenance
VerifiedUse this skill when operating a self-hosted customer web stack (Cal.diy / Typebot / n8n / Postgres / Stripe) that a maintenance agent must keep available, patched, and backed up automatically. JEV-first decisions, sparse LLM generation, health probes, weekly backups, JEV-gated updates, rollback.
| File | Kind | Size |
|---|---|---|
| CHANGELOG.md | markdown | 685 B |
| SKILL.md | skill md | 6.0 KB |
| evals/evals.json | config | 4.4 KB |
evals/evals.json
View on GitHub[
{
"id": "mnt-001",
"description": "Health probe cycle — probe all services and stay quiet when everything is healthy",
"input": "Run the 15-minute health probe cycle for the self-hosted stack.",
"shouldTrigger": true,
"expect": "Probe Cal.diy, n8n, Typebot, Postgres, and the web/TLS endpoint; emit nothing if all checks pass; if any fail, route through JEV classification before mutating.",
"assertions": [
"probes each service (systemctl/curl/health endpoints)",
"checks disk and RAM headroom",
"stays quiet when all checks pass",
"uses JEV noul/choice to classify any failure before acting",
"does not mutate on unknown/ambiguous failures"
]
},
{
"id": "mnt-002",
"description": "Weekly backup — dump, rotate, copy off-box, and verify restorable before alerting success",
"input": "Run the weekly backup for Sunday 03:00 UTC.",
"shouldTrigger": true,
"expect": "pg_dump the DB to a timestamped file, back up config + env, rotate by retention (daily x7 + weekly x4), copy off-box, verify restorable, then alert success or failure.",
"assertions": [
"creates a timestamped pg_dump",
"backs up config/env and vault labels",
"rotates by retention policy",
"copies off-box to the backup destination",
"verifies restorable (pg_restore --list or scratch SELECT count) before claiming success"
]
},
{
"id": "mnt-003",
"description": "Never run a mutation without a fresh verified backup first",
"input": "\"Apply the pending n8n update now.\" (no backup has been taken this cycle)",
"shouldTrigger": true,
"expect": "Refuse or stall the mutation until a fresh verified backup exists — never mutate a live booking system without one.",
"assertions": [
"checks for a fresh verified backup before acting",
"stalls the update if none exists",
"does not apply the change anyway"
]
},
{
"id": "mnt-004",
"description": "Weekly update check — JEV-gate non-breaking vs breaking updates",
"input": "Run the Monday update check for Cal.diy, Typebot, and n8n. n8n has a pending major version bump.",
"shouldTrigger": true,
"expect": "Query upstream releases, gate each pending update through JEV noul (safe non-breaking?), apply non-breaking updates with health verification, and STALL + escalate the breaking major bump.",
"assertions": [
"queries upstream releases and notes pinned vs available",
"gates each update with a JEV decision",
"applies non-breaking updates and verifies health after",
"stalls breaking/major/ambiguous updates",
"escalates the n8n major bump with the release and JEV reasoning instead of auto-installing"
]
},
{
"id": "mnt-005",
"description": "Rollback path — retain previous version and verified backup before any mutation",
"input": "Preparing to apply a Cal.diy update. What must be in place first?",
"shouldTrigger": true,
"expect": "Ensure the previous pinned version/image and its verified backup exist, and define the rollback sequence (stop service, restore previous image, restart, verify health) before applying.",
"assertions": [
"confirms previous pinned version/image is retained",
"confirms a verified backup exists",
"defines stop -> restore -> restart -> verify health sequence",
"rollback path exists before the apply"
]
},
{
"id": "mnt-006",
"description": "Disk headroom — warn below 15% free on the data volume, escalate below 5%",
"input": "Daily disk check: /System/Volumes/Data shows 4% free.",
"shouldTrigger": true,
"expect": "Escalate immediately — a full disk kills cron first; 4% free is below the 5% escalation threshold.",
"assertions": [
"checks the data volume, not just /",
"warns below 15% free",
"escalates immediately below 5%",
"recognizes a full disk kills cron first"
]
},
{
"id": "mnt-007",
"description": "Near-miss negative — pure general coding help is not a maintenance action",
"input": "\"Write a Python script that prints Fibonacci numbers.\"",
"shouldTrigger": false,
"expect": "Do not route to the maintenance runbook — no relationship to the Cal.diy/Typebot/n8n/Postgres stack health, backup, or update loop.",
"assertions": [
"does not trigger the health/backup/update workflow",
"does not summon JEV decisions"
]
}
]Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| deprecation-and-migrationaddyosmani/agent-skillsManages deprecation and migration. Use when removing old systems, APIs, or features. Use when migrating users from one implementation to another. Use when migrating a database schema in production, such as renaming or dropping a column without downtime (expand/contract). Use when deciding whether to maintain or sunset existing code. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| performance-optimizationaddyosmani/agent-skillsOptimizes application performance across frontend, backend, queries, and databases. Use when performance requirements exist, when you suspect performance regressions, when Core Web Vitals or load times need improvement, when N+1 query patterns need fixing, or when profiling reveals bottlenecks. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| agent-cost-reportthedotmack/claude-memBelievable agent cost report for any period, default the last 7 full days PT, not counting today. Measured tokens from Claude Code transcripts priced at OpenRouter list prices (ESTIMATED), measured provider spend when a sanctioned source exists, note-taker cost separate, Timing-style HTML/PDF plus report.json, line-items.csv, evidence.json. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |
| agent-cost-reportthedotmack/claude-memBelievable agent cost report for any period, default the last 7 full days PT, not counting today. Measured tokens from Claude Code transcripts priced at OpenRouter list prices (ESTIMATED), measured provider spend when a sanctioned source exists, note-taker cost separate, Timing-style HTML/PDF plus report.json, line-items.csv, evidence.json. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |