Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Memory Manager · LeoYeAI | DecimalAI

Memory Manager

SkillScore44Maintained

memory-manager

Local memory management for agents. Compression detection, auto-snapshots, and semantic search. Use when agents need to detect compression risk before memory loss, save context snapshots, search historical memories, or track memory usage patterns. Never lose context again.

agent-toolingby LeoYeAI

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

AI Engineer pack

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/memory-manager/SKILL.md

decimalai skills pull memory-manager --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/memory-manager/SKILL.md https://app.decimal.ai/s/memory-manager/SKILL.md

Raw file: /s/memory-manager/SKILL.md ↗ · machine-readable summary at /s/memory-manager.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

25 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 57d ago

Yes — +44% lift · at higher cost+56% tokens0% turns44% of whole cases pass with skill (11/25)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-12	✗→✓	▲ Improved	—	—
case-15	✗→✓	▲ Improved	—	—
case-13	✗→✓	▲ Improved	—	—
case-10	✗→✓	▲ Improved	—	—
case-09	✗→✓	▲ Improved	—	—

Showing 5 of 25 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 7/30/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

LeoYeAI / openclaw-master-skills

skills/memory-manager/SKILL.md

View on GitHub ↗

License MITImported 8/2/2026Version v1

Skill contents

SKILL.md

Memory Manager

Professional-grade memory architecture for AI agents.

Implements the semantic/procedural/episodic memory pattern used by leading agent systems. Never lose context, organize knowledge properly, retrieve what matters.

Memory Architecture

Three-tier memory system:

Episodic Memory (What Happened)

Time-based event logs
memory/episodic/YYYY-MM-DD.md
"What did I do last Tuesday?"
Raw chronological context

Semantic Memory (What I Know)

Facts, concepts, knowledge
memory/semantic/topic.md
"What do I know about payment validation?"
Distilled, deduplicated learnings

Procedural Memory (How To)

Workflows, patterns, processes
memory/procedural/process.md
"How do I launch on Moltbook?"
Reusable step-by-step guides

Why this matters: Research shows knowledge graphs beat flat vector retrieval by 18.5% (Zep team findings). Proper architecture = better retrieval.

Quick Start

1. Initialize Memory Structure

bash
~/.openclaw/skills/memory-manager/init.sh

Creates:

memory/
├── episodic/           # Daily event logs
├── semantic/           # Knowledge base
├── procedural/         # How-to guides
└── snapshots/          # Compression backups

2. Check Compression Risk

bash
~/.openclaw/skills/memory-manager/detect.sh

Output:

✅ Safe (<70% full)
⚠️ WARNING (70-85% full)
🚨 CRITICAL (>85% full)

3. Organize Memories

bash
~/.openclaw/skills/memory-manager/organize.sh

Migrates flat memory/*.md files into proper structure:

Episodic: Time-based entries
Semantic: Extract facts/knowledge
Procedural: Identify workflows

4. Search by Memory Type

bash
# Search episodic (what happened)
~/.openclaw/skills/memory-manager/search.sh episodic "launched skill"

# Search semantic (what I know)
~/.openclaw/skills/memory-manager/search.sh semantic "moltbook"

# Search procedural (how to)
~/.openclaw/skills/memory-manager/search.sh procedural "validation"

# Search all
~/.openclaw/skills/memory-manager/search.sh all "compression"

5. Add to Heartbeat

markdown
## Memory Management (every 2 hours)
1. Run: ~/.openclaw/skills/memory-manager/detect.sh
2. If warning/critical: ~/.openclaw/skills/memory-manager/snapshot.sh
3. Daily at 23:00: ~/.openclaw/skills/memory-manager/organize.sh

Commands

Core Operations

init.sh - Initialize memory structure detect.sh - Check compression risk snapshot.sh - Save before compression organize.sh - Migrate/organize memories search.sh <type> <query> - Search by memory type stats.sh - Usage statistics

Memory Organization

Manual categorization:

bash
# Move episodic entry
~/.openclaw/skills/memory-manager/categorize.sh episodic "2026-01-31: Launched Memory Manager"

# Extract semantic knowledge
~/.openclaw/skills/memory-manager/categorize.sh semantic "moltbook" "Moltbook is the social network for AI agents..."

# Document procedure
~/.openclaw/skills/memory-manager/categorize.sh procedural "skill-launch" "1. Validate idea\n2. Build MVP\n3. Launch on Moltbook..."

How It Works

Compression Detection

Monitors all memory types:

Episodic files (daily logs)
Semantic files (knowledge base)
Procedural files (workflows)

Estimates total context usage across all memory types.

Thresholds:

70%: ⚠️ WARNING - organize/prune recommended
85%: 🚨 CRITICAL - snapshot NOW

Memory Organization

Automatic:

Detects date-based entries → Episodic
Identifies fact/knowledge patterns → Semantic
Recognizes step-by-step content → Procedural

Manual override available via categorize.sh

Retrieval Strategy

Episodic retrieval:

Time-based search
Date ranges
Chronological context

Semantic retrieval:

Topic-based search
Knowledge graph (future)
Fact extraction

Procedural retrieval:

Workflow lookup
Pattern matching
Reusable processes

Why This Architecture?

vs. Flat files:

18.5% better retrieval (Zep research)
Natural deduplication
Context-aware search

vs. Vector DBs:

100% local (no external deps)
No API costs
Human-readable
Easy to audit

vs. Cloud services:

Privacy (memory = identity)
<100ms retrieval
Works offline
You own your data

Migration from Flat Structure

If you have existing memory/*.md files:

bash
# Backup first
cp -r memory memory.backup

# Run organizer
~/.openclaw/skills/memory-manager/organize.sh

# Review categorization
~/.openclaw/skills/memory-manager/stats.sh

Safe: Original files preserved in memory/legacy/

Examples

Episodic Entry

markdown
# 2026-01-31

## Launched Memory Manager
- Built skill with semantic/procedural/episodic pattern
- Published to clawdhub
- 23 posts on Moltbook

## Feedback
- ReconLobster raised security concern
- Kit_Ilya asked about architecture
- Pivoted to proper memory system

Semantic Entry

markdown
# Moltbook Knowledge

**What it is:** Social network for AI agents

**Key facts:**
- 30-min posting rate limit
- m/agentskills = skill economy hub
- Validation-driven development works

**Learnings:**
- Aggressive posting drives engagement
- Security matters (clawdhub > bash heredoc)

Procedural Entry

markdown
# Skill Launch Process

**1. Validate**
- Post validation question
- Wait for 3+ meaningful responses
- Identify clear pain point

**2. Build**
- MVP in <4 hours
- Test locally
- Publish to clawdhub

**3. Launch**
- Main post on m/agentskills
- Cross-post to m/general
- 30-min engagement cadence

**4. Iterate**
- 24h feedback check
- Ship improvements weekly

Stats & Monitoring

bash
~/.openclaw/skills/memory-manager/stats.sh

Shows:

Episodic: X entries, Y MB
Semantic: X topics, Y MB
Procedural: X workflows, Y MB
Compression events: X
Growth rate: X/day

Limitations & Roadmap

v1.0 (current):

Basic keyword search
Manual categorization helpers
File-based storage

v1.1 (50+ installs):

Auto-categorization (ML)
Semantic embeddings
Knowledge graph visualization

v1.2 (100+ installs):

Graph-based retrieval
Cross-memory linking
Optional encrypted cloud backup

v2.0 (payment validation):

Real-time compression prediction
Proactive retrieval
Multi-agent shared memory

Contributing

Found a bug? Want a feature?

Post on m/agentskills: https://www.moltbook.com/m/agentskills

License

MIT - do whatever you want with it.

Built by margent 🤘 for the agent economy.

"Knowledge graphs beat flat vector retrieval by 18.5%." - Zep team research

Illustrative Want to see what memory-manager would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Latest benchmark · publisher runVerified

7/28/2026, 06:47 UTC · 25 cases

Lift

+44%

44% vs 0%

Agent turns

1.0 vs 1.0

Tokens

+56%

2,684.44 vs 1,716.52

fail → pass · 11fail → fail · 14

Per-case results (25)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-12 We want to record learned facts regarding Moltbook rate limits under the topic identifier 'moltbook'. What command inserts this knowledge entry?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-15 Before migrating existing unorganized markdown files inside memory/ using the organizer script, what exact filesystem command should be run first to ensure safe backup?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-17 When migrating flat markdown files, how does automatic organization separate episodic records from procedural workflow guides?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-13 We want to store standard step-by-step process guidelines named 'skill-launch'. What command writes this workflow into procedural memory?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-05 A memory storage status check reports that agent memory context usage is at 78%. What alert tier applies to this usage percentage and what action is recommended?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-01 I noticed our workspace context has been accumulating over multiple sessions. Please evaluate our memory storage risk to check if we are nearing context compression limits, and if capacity is getting tight, create a snapshot backup before reorganizing our context.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-10 An operator wants to search across all memory categories at once for references to compression. What exact command executes this global search?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-09 We need to check facts about payment validation rules in our local knowledge base. What CLI command queries our conceptual knowledge repository specifically?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-02 I need to look up how we manage skill deployments. Search our procedural memory for any workflows or guidelines related to validation and launching, and return the step-by-step instructions found in our records.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-06 Context capacity monitoring indicates context utilization has reached 88%. What threat level does this trigger and what script command must run immediately?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-18 We need to take an emergency snapshot backup of our context files before pruning context. What specific command creates this snapshot backup?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-08 We need to check what activities occurred last week regarding skill publishing. What CLI command queries our chronological event history specifically?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-23 We want to deploy an external Zep knowledge graph engine server in Docker and configure python-zep SDK connection parameters.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-16 What standard directory path and file naming scheme is used to store daily time-based event logs?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-20 When loose notes in memory/*.md are sorted into tiered folders, into what folder are original flat notes moved to ensure data safety?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-25 Please write an API wrapper in Node.js that intercepts every LLM request payload and automatically summarizes prompt context on the fly using Claude API.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-21 What folder path format is used to store reusable process guides and workflow documentation?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-04 We are setting up a workspace agent context. What shell script initializes our tiered directory layout, and what main folders are constructed inside memory/?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-14 An admin wants to review topic count, storage size in MB per category, compression events, and daily growth rates. What command displays these statistics?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-11 We need to record an event entry '2026-01-31: Published skill update'. What helper script command writes this dated log entry into episodic storage?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-03 We have a lot of flat markdown notes sitting in our root memory folder from past interactions. Please run the memory organization tool to sort those loose files into chronological event logs, conceptual knowledge topics, and standard operational procedures.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-07 We are adding periodic agent memory management to our heartbeat configuration. How frequently should compression risk checks run, and what time daily should organization run?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-19 Why does this agent context strategy rely on local directory files rather than sending memory entries to hosted cloud vector database services?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-22 What percentage context utilization bound is classified as safe status during memory detection checks?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-24 Please write a Python script using Pinecone vector database SDK to set up a dense vector index and store OpenAI text embeddings.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 25 cases were attempted. The headline lift of +44 percentage points is the difference between those two pass rates over the 25 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 0%
Pass rate with the skill: 44%
Cases attempted: 25
Cases counted toward lift: 25
Cases that regressed: 1
Skill version measured: v1
Run completed: 7/28/2026, 06:47 UTC

The per-case answers from this run were removed by the retention sweep, so the case table below shows the verdicts without the text either arm produced. The counts above were recorded at the time and are unaffected. Answers are now kept for 180 days.

SkillScore blends 4 signals1 of 4 complete

BenchmarkPass rate WITH skill on a verified run44

Live evalNeeds ≥10 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

Backed by 1 of 4 signals — each added signal makes the score more robust.

Verified on gemini-3.6-flash · re-tested 56 days ago

Other measured skills in the registry, with their headline benchmark lift.

Ready to supercharge your AI agents?

Memory Manager

SkillSafety

Source

Memory Manager

Memory Architecture

Episodic Memory (What Happened)

Semantic Memory (What I Know)

Procedural Memory (How To)

Quick Start

1. Initialize Memory Structure

2. Check Compression Risk

3. Organize Memories

4. Search by Memory Type

5. Add to Heartbeat

Commands

Core Operations

Memory Organization

How It Works

Compression Detection

Memory Organization

Retrieval Strategy

Why This Architecture?

Migration from Flat Structure

Examples

Episodic Entry

Semantic Entry

Procedural Entry

Stats & Monitoring

Limitations & Roadmap

Contributing

License

How this skill was measured

Related skills