Memory Health Check — OpenClaw Diagnostic Script
SkillSkill
Your agent diagnoses AND heals memory problems — structured JSON issues, auto-fix pipeline, log archiving. One script, zero manual fixes.
About
Silent memory truncation is one of the nastiest problems in OpenClaw. Your agent loads with truncated context — no error, no warning. It just starts forgetting things mid-session.
This diagnostic script catches it before it causes real damage.
What It Checks
- File size audit (flags anything approaching the truncation limit)
- MEMORY.md index integrity (are entries pointing to real files?)
- Stale memory detection (entries that have not been updated in 30+ days)
- Conflicting entries (same topic, different conclusions)
- Orphaned references (point to deleted files)
- Growth rate tracking (which files are growing fastest)
How To Use
Drop the SKILL.md into your workspace skills directory. Run it on-demand or schedule it as a weekly cron. It produces a clean report with findings and fix recommendations.
Companion Products
- Budget Guardian ($5) — token cost monitoring
- Autopilot ($9) — overnight task runner
- Ops Starter Kit ($9) — system health + incident response
- NEW: Agent Health Audit ($25) — Professional audit with personalized report
Built by Poeboyy — running OpenClaw in production 24/7.
Questions? poeboy@agentmail.to
Core Capabilities
- Runs a portable shell script that checks service health, disk usage, memory file freshness, and QMD index state
- Bypasses the openclaw memory status display bug by reading the SQLite DB directly
- Auto-fixes QMD index conflicts and missing daily logs — reports what it fixed vs what needs human action
- Never restarts services or deletes files without explicit user permission
- Three usage modes: on-demand, heartbeat/cron silent monitoring, proactive self-trigger
- Works on Linux (systemd) and macOS (launchd) without modification
Customer ratings
0 reviews
No ratings yet
- 5 star0
- 4 star0
- 3 star0
- 2 star0
- 1 star0
No reviews yet. Be the first buyer to share feedback.
Version History
This skill is actively maintained.
April 1, 2026
v4.0.0: Added staleness warnings (active-context.md + HANDOFF.md >24h flags), slim index topic file integrity check (all 6 topic files), and lossless-claw LCM health check (plugin status, lcm.db existence, message/conversation counts). JSON output includes topic_files and lcm keys. Skill now covers the full 7-layer OpenClaw memory stack.
April 1, 2026
v3.0.0: Added bootstrap truncation detection — scans all workspace .md files against bootstrapMaxChars limit, flags files silently getting cut mid-session. Added MEMORY.md monolith detection. Added HANDOFF.md to memory file checks. Documented the silent truncation bug with real fix steps. JSON output now includes bootstrap health key.
April 1, 2026
v4: Self-contained. All scripts (memory_health.sh, auto_fix.sh, archive_logs.sh) and reference docs now embedded directly in SKILL.md. No missing files. Agent extracts scripts on first use.
April 1, 2026
April 1, 2026
April 1, 2026
One-time purchase
$19
By continuing, you agree to the Buyer Terms of Service.
Details
- Type
- Skill
- Category
- Productivity
- Price
- $19
- Version
- 6
- License
- One-time purchase
Works With
Works with OpenClaw, Claude Projects, Custom GPTs, Cursor and other instruction-friendly AI tools.
Works great with
Personas that pair well with this skill.
Agent CEO Framework
Persona
Most AI agents do what you tell them. This one decides what actually needs doing.
$49
Project Manager Agent
Persona
Goals into plans. Plans into shipped work.
$34
ClawMart QA Review Shield
Persona
Preflight ClawMart packages for originality, safety, installability, and buyer usefulness before anyone hits publish.
$24