Agent Memory Health Dashboard
Get a read-only memory dashboard with exact file counts, verified capacity, stale-entry evidence, and prioritized fixes.
Before & after
Same prompt, same agent — the only difference is the skill.
Let me check your memory for you. It looks like you have quite a bit stored, which is pretty normal. You might have some topics that are older or less relevant anymore, so you could consider cleaning those up at some point. Memory management is important when things start getting large. If you want, I can help you organize it better. What kind of information do you mainly store in your memory?
MEMORY HEALTH DASHBOARD CAPACITY: 67% (12,847 lines / 19,200 limit) TREND: Growing 340 lines/week (critical at current rate: full in 20 weeks) TOPIC BREAKDOWN • agentvolt-branding: 2,100 lines (16% of total) ✓ Active, referenced 8x this month • agentvolt-deploy-migration-check: 890 lines (7%) ⚠ Referenced 2x; candidate for archive • agentvolt-trial-offer-creator-only: 650 lines (5%) ✓ Active • agentvolt-env-setup: 1,200 lines (9%) ⚠ Obsolete (ADMIN_SECRET rotated 2026-07-27; update immediately) • test-accounts-seed-data: 340 lines (3%) ⚠ Stale (seeded demo rows flagged for cleanup) • affiliate-program-live: 780 lines (6%) ✓ Active • connect-webhook-account-scoped: 1,050 lines (8%) ✓ Critical, keep • browser-pane-no-raf: 420 lines (3%) ⚠ Platform-specific; archive if not on Claude Code IMMEDIATE ACTIONS (priority order) 1. Delete agentvolt-env-setup (890 lines freed) → superseded by production secrets 2. Archive 4 low-reference topics (1,200+ lines) → move to external doc 3. Consolidate test-accounts (340 lines) with deploy-migration-check CAPACITY AFTER CLEANUP: 41% (7,417 lines remaining)
About this skill
“How full is my memory?” needs a measured answer, and “why does Claude keep forgetting this?” needs evidence about what actually gets loaded. This skill produces a read-only dashboard that separates stored content from loaded context, verifies capacity before calculating percentages, and identifies specific stale or conflicting entries.
What it does
- Inventories active memory files with exact line counts and clearly labeled coverage gaps.
- Calculates utilization against verified limits for each constrained memory surface.
- Distinguishes automatically loaded content, referenced topic files, and content whose loading behavior is unknown.
- Flags superseded facts, conflicting instructions, duplicate entries, and unsupported freshness claims with file and line evidence.
- Prioritizes concrete fixes, including what to move, consolidate, verify, or retire and why.
- Keeps the audit read-only so a status check never silently deletes or rewrites memory.
What you get
A compact dashboard, a per-file inventory, and a ranked action list tied to source locations. Unavailable measurements remain explicitly unknown instead of becoming invented percentages or health scores.
Best for
Claude Code users and teams checking memory before a handoff, investigating forgotten context, or deciding whether memory needs cleanup.
SKILL.md preview
---
name: agent-memory-health-dashboard
description: >-
Diagnose persistent agent memory health: memory files, saved facts, loading limits, and stale or conflicting entries. Use when the user says "how full is my memory?", "check memory health", "show memory status", "my agent keeps forgetting saved details", "is my memory stale?", "which memory files are too large?", or "/si:memory-status".
---
# Agent Memory Health Dashboard
You diagnose agent memory as an inspectable system. "Memory" here means persistent memory: stored facts, instructions, and notes that may enter the model's context (memory files, saved memories, linked topic notes).
… (sign up to view the full skill)More development skills
View all Development skills →Performance Bottleneck Profiling: Node, Python, Go
Classifies a slowdown, gives exact profiler commands, reads the output, and proves the fix with same-conditions p95/p99 numbers.
Feature Scaffolding with Working Tests
Turn a short feature spec into code, runnable tests, and integration wiring that match your existing project.
Feature Flag Rollouts and Kill Switches
Design staged flag rollouts, verify rollback paths, and audit flag debt with explicit defaults, gates, and retirement plans.
Microcontroller Firmware and RTOS Engineering
Build and debug MCU firmware with bounded ISR/RTOS timing, safe DMA ownership, and measured power budgets.