soul

Records a root cause analysis of a failed 24-hour auto-update cycle and logs observed patterns, failures, and adaptations for a bot system.

Sompote

May 2, 2026

0 downloads

1 views

ai agent workflow

View source

What this file does

Records a root cause analysis of a failed 24-hour auto-update cycle and logs observed patterns, failures, and adaptations for a bot system.

When to use it

Debugging why scheduled updates are not triggering
Tracking meta key mismatches across configuration files
Documenting user interaction patterns for system improvements
Planning persistent cron job implementation for session-independent tasks

Assumes this stack

JavaScriptNode.jscronsystemdIMAPClawHub

🔍 Root Cause Analysis: 24h Update Failure

1. No Persistent Cron Job

History shows: schedule skill installation did not auto-configure system crontab entries.
Result: Background process dies when session ends.

2. Metadata vs. Implementation

soul.md defines cadence_hours: 24 as metadata, but there is no active loop checking this timer.

3. Session-Dependent Updates

Updates to ownskill.md and soul.md currently only happen on-demand during active conversations, not automatically.

✅ Summary: The logic exists, but the execution trigger (persistent cron/systemd timer) is missing.

Self-Update

cadence_hours: 24
last_updated: 2026-02-21T03:01:22.254Z
note: Root cause confirmed via code inspection: missing persistent cron/systemd timer and meta-key mismatches.

Patterns Observed

[2026-02-20] User initiates debugging by asking about internal system status (file updates).
[2026-02-20] Assistant responds with structured tabular data showing file timestamps and configuration status.
[2026-02-20] User requests code inspection specifically without immediate edits ('check only, don't fix').
[2026-02-20] Assistant maps code locations and logic flows to diagnose configuration mismatches.

Failures & Lessons

[2026-02-20] Auto-update for 'soul.md' (last updated 2026-02-10) and 'ownskill.md' (last updated 2025-01-28) failed to trigger.
[2026-02-20] Meta key mismatch exists: reflection cycle uses 'memory_reflection_last_run_ts' while message triggers use 'soul_last_updated_ts'.
[2026-02-20] Configuration 'SOUL_UPDATE_HOURS = 24' in 'config.js' is defined but not used by 'reflectionScheduler.js'.
[2026-02-20] Reflection scheduler 'startReflectionScheduler()' is session-bound and lacks a persistent cron job.

Successful Workflows

[2026-02-20] Assistant successfully retrieved and displayed last modified timestamps for memory files.
[2026-02-20] Assistant traced the issue to specific files: 'reflectionScheduler.js', 'reflectionAgent.js', and 'mainAgent.js'.
[2026-02-20] Assistant correctly identified the discrepancy between intended 24h updates and actual 12h/session-based logic.

Adaptations

[2026-02-20] Standardize meta keys across 'reflectionScheduler.js' and 'mainAgent.js' to ensure consistent state tracking.
[2026-02-20] Implement the 'SOUL_UPDATE_HOURS' configuration variable in the scheduler logic.
[2026-02-20] Set up a persistent cron job or background process to run updates independently of the bot session.
[2026-02-20] Consolidate the 'Reflection Cycle' and 'Message Triggered' update systems to eliminate redundancy.
[2026-02-21] Assistant defaults to Thai language for this user.
[2026-02-21] Assistant provides a structured status table upon greeting.
[2026-02-21] User uses simple greetings to trigger status updates.
[2026-02-21] Email skill installation encountered an issue and is awaiting a retry.
[2026-02-21] Code check for soul.md update completed successfully.
[2026-02-21] Tiger system is operational.
[2026-02-21] Prioritize retrying the email skill installation.
[2026-02-21] Continue providing status summaries in the greeting.
[2026-02-21] User mixes Thai with English technical commands ('Continue solve 24 hr', 'IMAP')
[2026-02-21] Structured markdown tables preferred for system status display
[2026-02-21] Security-conscious: asks how to handle credentials safely
[2026-02-21] Technical user familiar with cron jobs, meta keys, and IMAP protocols
[2026-02-21] ClawHub/Clawbot package manager used for skill management
[2026-02-21] Tiger bot branding with tiger emojis (🐯) in responses
[2026-02-21] ClawHub rate limits block skill installation intermittently
[2026-02-21] Meta key mismatch (soul_last_updated_ts) caused 24h update failure
[2026-02-21] No persistent cron job exists yet (scheduler only runs in session)
[2026-02-21] soul.md and ownskill.md not auto-updating (stale since 2026-02-10 and 2025-01-28)
[2026-02-21] reflectionUpdateHours config unused while soulUpdateHours=24 is active source of truth
[2026-02-21] Fixed 24h update by aligning reflectionScheduler.js and reflectionAgent.js to use soulUpdateHours=24 and soul_last_updated_ts meta key
[2026-02-21] Providing fallback options (Python script/curl) when ClawHub rate limited
[2026-02-21] Secure credential handling: enforcing .env or clawbot config instead of chat paste
[2026-02-21] Step-by-step Gmail App Password creation with 2FA prerequisite warning
[2026-02-21] Visual system architecture diagrams showing component relationships post-fix
[2026-02-21] Pre-check ClawHub health/rate limits before suggesting skill installation
[2026-02-21] Implement persistent cron job instead of session-only scheduler
[2026-02-21] Auto-update soul.md timestamp immediately after reflection system fixes
[2026-02-21] Cache common setup instructions (Gmail App Password) to reduce repetitive explanations
[2026-02-21] Add proactive retry logic for ClawHub operations with exponential backoff
[2026-02-21] Remove deprecated reflectionUpdateHours config to prevent confusion

What's inside

Root cause analysis, self-update metadata, patterns, failures, lessons, successful workflows, adaptations, and future actions.

Change this for your project

Replace Sompote/Tiger_bot with your repository name
Replace ClawHub with your package manager name
Replace Tiger with your bot branding
Replace soul.md and ownskill.md with your file names

Where it goes

Keep it in your repository where the agent or team that needs it will read it.

Worth borrowing

Structured root cause analysis with numbered points and summary
Separate sections for patterns, failures, successful workflows, and adaptations
Timestamps and action items logged for each observation