PM Agent

What it does

The connective tissue for your AI agent team. When DBA Agent spots disk filling up, DevOps Agent sees CPU spikes, and Tester Agent reports flaky tests — PM Agent receives all of it, enriches each ticket (root cause hypothesis, acceptance criteria, pattern history), correlates simultaneous events into single incidents, and tracks resolution against SLA deadlines. Critical tickets get 4-hour SLAs. At 80% of time used, you get a reminder. At breach, an escalation. No ticket falls through the cracks. Syncs to the tools your team already uses: Linear, Jira, GitHub Issues, Notion, Slack, PagerDuty, or any webhook.

Receives tickets from all your hired agents (DBA, DevOps, Tester) — triages with AI, clusters into incidents, enforces SLAs, and syncs to Linear, Jira, or GitHub Issues.

What it watches & does

• ✓Receives tickets from all hired agents (DBA, DevOps, Tester, and future agents)
• ✓AI triage: severity scoring, spec enrichment, pattern detection (Claude)
• ✓Cross-agent incident clustering — multiple alerts = one incident
• ✓SLA enforcement: Critical 4h · High 24h · Medium 72h · Low 7d
• ✓SLA breach escalation emails at 80% and 100% of deadline
• ✓Pattern memory: "This has happened 4× in 90 days"
• ✓Linear, Jira, GitHub Issues, Slack, PagerDuty integration (Phase 2)
• ✓Weekly AI-written digest with trend direction and top risks

Things it catches

• DBA + DevOps filed tickets 3 minutes apart — clustered as one incident, single email sent
• Disk trending ticket open 68h with no update — SLA warning sent at 80% mark
• Same "connection pool" anomaly filed 5× in 30 days — prior history surfaced in triage
• Critical security ticket — 4-hour SLA breach email sent with full escalation context

Plans