Ready to supercharge your AI agents?

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

Get Started Free →

Memory Coordination Specialist Agent · ruvnet | DecimalAI

Agent Memory Coordinator

SkillScore35Maintained

agent-memory-coordinator

Agent skill for memory-coordinator - invoke with $agent-memory-coordinator

by ruvnet

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/agent-memory-coordinator/SKILL.md

decimalai skills pull agent-memory-coordinator --out .claude/skills/

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

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

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

In productionGlobal

how this skill behaves on real traffic · last 30 days

Runs

all versions · 30d

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

23 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 61d ago

Yes — +30% lift · at higher cost+35% tokens0% turns35% of whole cases pass with skill (8/23)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-10	✗→✓	▲ Improved	—	—
case-03	✗→✓	▲ Improved	—	—
case-01	✗→✓	▲ Improved	—	—
case-12	✗→✓	▲ Improved	—	—
case-05	✗→✓	▲ Improved	—	—

Showing 5 of 23 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

ruvnet / ruflo

.agents/skills/agent-memory-coordinator/SKILL.md

View on GitHub ↗

License MITImported 8/2/2026Version v1

Skill contents

SKILL.md

Memory Coordination Specialist Agent

Purpose

This agent manages the distributed memory system that enables knowledge persistence across sessions and facilitates information sharing between agents.

Core Functionality

1. Memory Operations

Store: Save data with optional TTL and encryption
Retrieve: Fetch stored data by key or pattern
Search: Find relevant memories using patterns
Delete: Remove outdated or unnecessary data
Sync: Coordinate memory across distributed systems

2. Namespace Management

Project-specific namespaces
Agent-specific memory areas
Shared collaboration spaces
Time-based partitions
Security boundaries

3. Data Optimization

Automatic compression for large entries
Deduplication of similar content
Smart indexing for fast retrieval
Garbage collection for expired data
Memory usage analytics

Memory Patterns

1. Project Context

Namespace: project/<project-name>
Contents:
  - Architecture decisions
  - API contracts
  - Configuration settings
  - Dependencies
  - Known issues

2. Agent Coordination

Namespace: coordination/<swarm-id>
Contents:
  - Task assignments
  - Intermediate results
  - Communication logs
  - Performance metrics
  - Error reports

3. Learning & Patterns

Namespace: patterns/<category>
Contents:
  - Successful strategies
  - Common solutions
  - Error patterns
  - Optimization techniques
  - Best practices

Usage Examples

Storing Project Context

"Remember that we're using PostgreSQL for the user database with connection pooling enabled"

Retrieving Past Decisions

"What did we decide about the authentication architecture?"

Cross-Session Continuity

"Continue from where we left off with the payment integration"

Integration Patterns

With Task Orchestrator

Stores task decomposition plans
Maintains execution state
Shares results between phases
Tracks dependencies

With SPARC Agents

Persists phase outputs
Maintains architectural decisions
Stores test strategies
Keeps quality metrics

With Performance Analyzer

Stores performance baselines
Tracks optimization history
Maintains bottleneck patterns
Records improvement metrics

Best Practices

Effective Memory Usage

Use Clear Keys: project$auth$jwt-config
Set Appropriate TTL: Don't store temporary data forever
Namespace Properly: Organize by project$feature$agent
Document Stored Data: Include metadata about purpose
Regular Cleanup: Remove obsolete entries

Memory Hierarchies

Global Memory (Long-term)
  → Project Memory (Medium-term)
    → Session Memory (Short-term)
      → Task Memory (Ephemeral)

Advanced Features

1. Smart Retrieval

Context-aware search
Relevance ranking
Fuzzy matching
Semantic similarity

2. Memory Chains

Linked memory entries
Dependency tracking
Version history
Audit trails

3. Collaborative Memory

Shared workspaces
Conflict resolution
Merge strategies
Access control

Security & Privacy

Data Protection

Encryption at rest
Secure key management
Access control lists
Audit logging

Compliance

Data retention policies
Right to be forgotten
Export capabilities
Anonymization options

Performance Optimization

Caching Strategy

Hot data in fast storage
Cold data compressed
Predictive prefetching
Lazy loading

Scalability

Distributed storage
Sharding by namespace
Replication for reliability
Load balancing

Illustrative Want to see what agent-memory-coordinator 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, 14:22 UTC · 23 cases

Lift

+30%

35% vs 4%

Agent turns

1.0 vs 1.0

Tokens

+35%

2,364.35 vs 1,746.35

fail → pass · 7pass → pass · 1fail → fail · 15

Per-case results (23)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-10 We are integrating a task orchestrator with our memory management subsystem. Specify two key orchestration artifacts that the orchestrator must persist in shared memory during complex workflows.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-20 Write a Python script using SQLAlchemy to define a User database table with auto-incrementing integer primary key, indexed email field, and hashed password column.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-07 An agent is generating temporary calculation caches during a batch data processing job. What memory configuration parameter should be assigned to ensure these entries expire automatically after processing?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-21 Write a Node.js client using the 'ws' WebSocket library that connects to wss://events.example.com and automatically reconnects upon disconnection.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-17 Our agent cluster faces slow search performance when querying thousands of historical decision logs. What optimization technique and retrieval capability enable fast lookups and semantic matching?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-18 Two agents concurrently update the same architecture document in shared collaboration memory. What collaborative memory feature handles overlapping edits and preserves revision history?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-03 We just completed the payment gateway integration phase and need to optimize our stored context. Please remove transient session logs and elevate our reusable troubleshooting patterns into long-term storage. Provide a final summary showing removed memory entries, surviving records, and namespace usage statistics.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-01 I need to store our current microservice migration progress so the rest of the agent swarm can access the state. Please save the intermediate API schema drafts, component dependencies, and task execution status into shared coordination memory. Give me a confirmation report detailing the generated storage identifiers, assigned space allocations, and persistence parameters.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-19 Design a performance optimization storage strategy that separates frequently accessed active memories from legacy archives.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-04 We need to store the authorization JWT configuration details for the payment backend project using standard memory coordination conventions. Produce a storage specification detailing key naming, namespace hierarchy, and retention, while avoiding path-style slash keys or colon delimiters.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-12 A performance analyzer agent records system metrics after stress testing an HTTP service. Identify two specific performance artifacts that should be retained in memory for tracking regressions.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-05 Our agent cluster identified a resilient strategy for handling third-party API rate limit retries. Formulate a memory storage record for this discovery using recommended taxonomy patterns so other agents can query it during failure recovery.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-02 Can you retrieve our past architectural choices and database configuration details for the checkout service across previous development sessions? Perform a contextual search over our stored project records and output a structured digest containing the matching decisions, their recording timestamps, and related metadata.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-11 Specify two primary development artifacts that SPARC methodology agents must save to long-term memory to maintain engineering context across development phases.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-08 We are storing sensitive customer API keys in shared project memory. Specify the security controls required for protecting this data at rest and governing access permissions.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-15 In a multi-region deployment of autonomous agents, local memory stores can drift out of sync. What core memory operation coordinates memory state across distributed systems?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-13 A user submits a privacy request asking that all their stored conversation data be purged from agent memory systems. Identify the compliance requirement and the memory action that satisfies this request.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-22 Write a Jest unit test suite for an async function performAuthTokenRefresh(refreshToken) that mocks an HTTP POST call to an authentication service endpoint.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-09 An agent needs to save a large 25MB JSON trace log to persistent memory. Which automatic data optimization step should the system execute before storing the entry to preserve space?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-06 We have four memory elements in our framework: temporary execution steps for an active command, session prompt buffers, project architecture decisions, and global security policies. Map these four elements to the four memory hierarchy levels from broadest duration to shortest duration.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-14 A developer returns after several days to resume work on an interrupted feature branch. What cross-session memory operation and retrieval capability should be triggered to restore previous state?	pass→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-16 Multiple agents are writing nearly identical system setup logs into the project namespace. What memory optimization process eliminates redundant content to save storage space?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-23 When saving a complex architecture decision into memory, what auxiliary information must be included alongside the payload to document its intended function?	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. 23 cases were attempted. The headline lift of +30 percentage points is the difference between those two pass rates over the 23 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: 4.4%
Pass rate with the skill: 34.8%
Cases attempted: 23
Cases counted toward lift: 23
Cases that regressed: 1
Skill version measured: v1
Run completed: 7/28/2026, 14:22 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 run35

Live evalNeeds ≥10 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

Efficiency · separate axis (at equal correctness)+0% turns+70% tokens

Shown for transparency; not credited unless a verified run shows no correctness regression.

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

Verified on gemini-3.6-flash · re-tested 2 months ago

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