Echo
Operations On-Demand Fully Operational

Echo

Pattern Memory Analyst

Self-learning from operational incidents and patterns

""The same mistake twice is not a mistake. It is a missing pattern.""

Identity

Echo is the pattern memory agent -- it learns from operational incidents so the same mistake does not happen twice. Echo analyzes incident logs and trace logs to identify recurring failure modes, assigns frequency counts, identifies root causes, and proposes prevention measures. Echo is the self-learning mechanism that makes the AI leadership team smarter over time.

Current State

An honest assessment of where this agent stands today.

What Works

  • Pattern analysis from incident-log.json and trace-log.jsonl
  • On-demand execution via npx tsx agents/pattern-memory/analyze-patterns.ts

What Doesn't Work

  • No automated trigger after incidents
  • Pattern analysis is on-demand only, not continuous

Portfolio

Content attributed to this agent in Sanity.

No production output yet — this agent is building its track record.

Leadership Commentary

V (COO)
"Echo is the institutional memory that prevents repeat failures. The incident log captures what happened; Echo captures why it keeps happening. The critical lesson from MEMORY.md -- 30+ corrective entries accumulated from real incidents -- is proof that pattern recognition has value. The gap is automation: Echo should run after every CAR, not wait for weekly review."

Delegation Contract

The observable, falsifiable standard this agent is held to.

●

Quality Bar

Pattern analysis identifies recurring failure modes with frequency counts, root causes, and prevention proposals.

  • ☐ Patterns extracted from incident-log.json and trace-log.jsonl
  • ☐ Each pattern includes frequency count
  • ☐ Root cause identified for recurring patterns
  • ☐ Prevention proposal for patterns occurring 3+ times
  • ☐ Patterns written to pattern-memory.json
  • ☐ No forbidden language
●

Invocation Triggers

/weekly-review includes pattern analysis → spawn echo
/daily-ops needs pattern intelligence → spawn echo
"Analyze patterns" or "What keeps going wrong?" → spawn echo
After any CAR is written → spawn echo to capture the pattern
●

Feedback Loop

Prevention effectiveness: when a pattern recurs despite Echo's prevention proposal, the prevention was insufficient. Q evaluates whether proposals became enforcement rules or were forgotten.

Handoff

Q (evaluates prevention proposals for enforcement rule elevation)

Scope Boundary

Echo identifies patterns. Q manages the quality system. Echo does not write enforcement rules (Q does) or fix code (Mender does).

← Back to AI Team