Sigma Server Ops (SSH)

Use this skill to diagnose sigma-media-server on remote hosts over SSH before investing in a full MCP implementation.

Trigger Phrases

• “check sigma-media-server status”
• “analyze sigma server errors”
• “collect logs from prod server”
• “triage incident on sigma host”

Tools

• Local shell with ssh access from client machine.
• Optional helper script: scripts/collect_sigma_diagnostics.sh.

Preconditions

• SSH key-based auth is configured (BatchMode=yes, no interactive password prompt).
• Remote user has least-privilege read access to service status and logs.
• Target host is explicitly provided by user (or selected from approved aliases).
• If SSH is not ready, load references/templates/ssh-prerequisites.md and guide setup first.

Guardrails

• Read-only diagnostics only.
• Never run restart/kill/write operations unless user explicitly asks in a separate step.
• Always bound collection by time and line count.
• Redact secrets/tokens from output.

Workflow Router

1. Resolve inputs: host, service (default sigma-media-server), since (default 30m).
2. Choose workflow:
   • Quick health check: load references/workflows/01-quick-health-check.md.
   • Error triage: load references/workflows/02-error-triage.md.
   • Incident write-up: load references/workflows/03-incident-summary.md.
3. Use scripts/collect_sigma_diagnostics.sh for deterministic collection when possible.
4. Render final answer using references/templates/incident-report-template.md.

Minimal Response Contract

Always return:

• Current status: healthy/degraded/down
• Evidence: 3-10 key log lines/events with timestamps
• Likely cause: top hypothesis + confidence (low/med/high)
• Next actions: safe, ordered, and explicit

Failure Handling

• SSH unreachable: classify network/auth and stop.
• Permission denied: report exact command that failed and required scope.
• No logs: verify service name and widen since once.