/agentem:status

Dashboard command that aggregates agent state, live GitHub data, and feedback scores into a single health snapshot.

Steps

1. Load agent state. Read context/agent-state.json if it exists. If it doesn't, note "No agent runs recorded — run an agent first or use /agentem:init to get started."

2. Agent Activity. For each agent in the state file, display:

   • Last run time (relative: "2h ago", "1d ago")
   • Total run count
   • Last summary (one-line description from most recent run)
   • If an agent has never run, show "—"

Agent Activity:
  risk-detector ........... 2h ago (5 runs) — Found 2 critical, 1 high, 3 medium risks
  spec-generator .......... 1d ago (2 runs) — Generated spec for notification preferences
  review-orchestrator ..... 3d ago (1 run)  — Routed 3 PRs, assigned 5 reviewers
  ticket-decomposer ....... —
  release-manager ......... —
  retro-analyzer .......... —

3. Current Risks. From the last risk-detector run in the state file:
   • List active signals grouped by severity (Critical/High/Medium/Low)
   • Show count of dismissed signals
   • If risk-detector has never run, show "No risk data — run risk-detector first"

Current Risks (from last scan 2h ago):
  Critical: 0
  High: 1 — PR #142 open 72h without review
  Medium: 3
  Low: 2
  Dismissed: 1

4. PR Review Status. Live query via gh pr list --json number,title,createdAt,reviewRequests,reviews,state:
   • Count open PRs
   • Count PRs awaiting review (no approvals)
   • Count PRs exceeding 48h without review
   • If gh is not available, note "gh CLI not available — install for live PR data"

PR Review Status (live):
  Open PRs: 7
  Awaiting review: 3
  Exceeding 48h SLA: 1 (PR #142)

5. Sprint Health. Composite assessment based on available data:
   • GREEN — No critical/high risks, no PRs exceeding SLA, agents running regularly
   • YELLOW — High risks present OR PRs approaching SLA OR agents haven't run in 3+ days
   • RED — Critical risks present OR multiple PRs exceeding SLA OR agents haven't run in 7+ days
   • Show the specific factors driving the assessment

Sprint Health: YELLOW
  ⚠ 1 PR exceeding 48h review SLA
  ⚠ ticket-decomposer hasn't run this sprint
  ✓ No critical risks
  ✓ risk-detector ran within 24h

6. Agent Effectiveness. From feedback data in the state file:
   • Show per-agent feedback scores (Yes/Partial/No counts)
   • Only show agents that have feedback data

Agent Effectiveness:
  risk-detector ........... 8 useful, 3 partial, 1 not useful (67% useful)
  spec-generator .......... 4 useful, 1 partial (80% useful)

7. Suggested Actions. Generate 1-3 contextual next steps based on the data above:
   • Prioritize by impact (critical risks > SLA breaches > missing agent runs > feedback improvements)
   • Each suggestion includes the specific command or action to take

Suggested Actions:
  1. Run review-orchestrator — PR #142 needs reviewer assignment (72h without review)
  2. Run ticket-decomposer — no tickets decomposed this sprint
  3. Fill in context/team/topology.md — improves review routing accuracy

Error Handling

• No state file → show minimal dashboard with "Run /agentem:init then run an agent to populate this dashboard"
• gh CLI not available → skip PR Review Status section, note the gap
• State file corrupted/invalid JSON → warn user, suggest deleting and re-running agents