Monitor Episodic Archiver

Monitor Episodic Archiver

Track session health, run nightly analysis with user profiling, and catch aging unresolved sessions.

Continuous Operation (Non-Negotiable)

This skill is always-on. It:

• Runs on its configured schedule indefinitely — it NEVER stops unless explicitly halted by the user
• The agent MUST NOT stop and wait for the human to ask for status or remember to check
• If a cycle fails, diagnose the failure, attempt auto-repair, and continue
• Only escalate to the human if genuinely blocked after exhausting /dogpile research
• Gracefully handles restarts and maintains state across cycles
• Is designed for multi-day/week/month autonomous operation

Anti-pattern: Reporting status and waiting for the human to ask "what next?" is UNACCEPTABLE. The agent must proactively fix issues and continue the monitoring loop.

Quick Start

cd .pi/skills/monitor-episodic-archiver

# Health dashboard
./run.sh dashboard

# Quick health check (JSON for automation)
./run.sh check --json

# List unresolved sessions with aging flags
./run.sh list-unresolved

# Analyze failure patterns
./run.sh analyze-patterns

# Run nightly pipeline (archive + analyze + profile)
./run.sh nightly --hours 24

# Dry run (see what would be processed)
./run.sh nightly --dry-run --json

# Register for daily scheduler
./run.sh register-nightly

Nightly Pipeline

3:00 AM daily (via scheduler)
    |
    v
Archive recent sessions (--no-analyze, fast)
    |
    v
Re-analyze with high-fidelity taxonomy (LLM)
    |
    v
Extract user behavioral profiles
    |
    v
Merge into user_priors (RGMem incremental)
    |
    v
Store lessons to /memory with bridge tags
    |
    v
Health check + report
    |
    v
~/.pi/monitor-episodic-archiver/nightly_report.json

Model Selection

• Default: deepseek-ai/DeepSeek-V3.1-TEE (0.60s latency, 6 instances)
• Override: ./run.sh nightly --model <model-id>
• Set via CHUTES_MODEL_ID env var

Commands

dashboard - Rich health overview

check - Automated health check (exit 0/1/2)

list-unresolved - Sessions needing attention

analyze-patterns - Failure pattern analysis

nightly - Full nightly pipeline

register-nightly - Register with scheduler

Health Criteria

Integration

State Files

~/.pi/monitor-episodic-archiver/
  health_report.json       # Latest health check
  nightly_report.json      # Latest nightly pipeline report
  nightly_state.json       # Nightly run history
  task_state_nightly.json  # Task-monitor progress tracking
  pattern_cache.json       # Cached failure patterns
  alert_history.jsonl      # Alert log

Environment Variables