Ready to supercharge your AI agents?

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

Memory Management Skill · ruvnet | DecimalAI

Memory Management

SkillScore65Maintained

memory-management

AgentDB memory system with HNSW vector search. Provides 150x-12,500x faster pattern retrieval, persistent storage, and semantic search capabilities for learning and knowledge management.

agent-toolingby ruvnet

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

Automatic

Knowledge Management pack

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

pip install decimalai

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

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

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

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

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

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 — +61% lift-7% tokens0% turns65% of whole cases pass with skill (15/23)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-10	✗→✓	▲ Improved	—	—
case-07	✗→✓	▲ Improved	—	—
case-13	✗→✓	▲ Improved	—	—
case-19	✗→✓	▲ Improved	—	—
case-11	✗→✓	▲ 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/memory-management/SKILL.md

View on GitHub ↗

License MITImported 8/2/2026Version v1Bundled 2 files · scripts

Skill contents

scripts/

SKILL.md

Memory Management Skill

Purpose

AgentDB memory system with HNSW vector search. Provides 150x-12,500x faster pattern retrieval, persistent storage, and semantic search capabilities for learning and knowledge management.

When to Trigger

need to store successful patterns
searching for similar solutions
semantic lookup of past work
learning from previous tasks
sharing knowledge between agents
building knowledge base

When to Skip

no learning needed
ephemeral one-off tasks
external data sources available
read-only exploration

Commands

Store Pattern

Store a pattern or knowledge item in memory

bash
npx @claude-flow/cli memory store --key "[key]" --value "[value]" --namespace patterns

Example:

bash
npx @claude-flow/cli memory store --key "auth-jwt-pattern" --value "JWT validation with refresh tokens" --namespace patterns

Semantic Search

Search memory using semantic similarity

bash
npx @claude-flow/cli memory search --query "[search terms]" --limit 10

Example:

bash
npx @claude-flow/cli memory search --query "authentication best practices" --limit 5

Retrieve Entry

Retrieve a specific memory entry by key

bash
npx @claude-flow/cli memory get --key "[key]" --namespace [namespace]

Example:

bash
npx @claude-flow/cli memory get --key "auth-jwt-pattern" --namespace patterns

List Entries

List all entries in a namespace

bash
npx @claude-flow/cli memory list --namespace [namespace]

Example:

bash
npx @claude-flow/cli memory list --namespace patterns --limit 20

Delete Entry

Delete a memory entry

bash
npx @claude-flow/cli memory delete --key "[key]" --namespace [namespace]

Initialize HNSW Index

Initialize HNSW vector search index

bash
npx @claude-flow/cli memory init --enable-hnsw

Memory Stats

Show memory usage statistics

bash
npx @claude-flow/cli memory stats

Export Memory

Export memory to JSON

bash
npx @claude-flow/cli memory export --output memory-backup.json

Scripts

| Script | Path | Description | |--------|------|-------------| | memory-backup | .agents/scripts/memory-backup.sh | Backup memory to external storage | | memory-consolidate | .agents/scripts/memory-consolidate.sh | Consolidate and optimize memory |

References

| Document | Path | Description | |----------|------|-------------| | HNSW Guide | docs/hnsw.md | HNSW vector search configuration | | Memory Schema | docs/memory-schema.md | Memory namespace and schema reference |

Best Practices

Check memory for existing patterns before starting
Use hierarchical topology for coordination
Store successful patterns after completion
Document any new learnings

Illustrative Want to see what memory-management 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, 14:49 UTC · 23 cases

Lift

+61%

65% vs 4%

Agent turns

1.0 vs 1.0

Tokens

-8%

1,090.74 vs 1,179

fail → pass · 14pass → pass · 1fail → fail · 8

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 Initialize vector search indexing capabilities for our memory store to improve search speed. Provide the CLI command.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-07 Fetch the exact saved entry associated with key 'auth-jwt-pattern' from our 'patterns' namespace. Output the CLI command to retrieve it.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-13 Which executable script in the repository performs memory consolidation and optimization?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-19 We created a database pooling pattern for PostgreSQL. Save this with key 'postgres-pool-config' and value 'pg.Pool max 20' in the 'patterns' namespace using the CLI.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-01 We just finalized a standard solution for handling resilient API retries with exponential backoff. Please save this architectural pattern into our persistent memory under the identifier 'api-retry-backoff' along with the execution instructions, so future tasks can reference it.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-11 Audit current agent memory usage and index status. What CLI command displays these memory metrics?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-08 List all stored memory entries under the 'patterns' namespace, displaying up to 20 items. Output the CLI command.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-22 Delete the memory record identified by key 'deprecated-logger' in the 'logging' namespace via CLI command.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-04 We need to fetch live currency exchange rates from an external REST service for a temporary math calculation. Output the HTTP command or script to make this external request.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-03 We are preparing for a system migration and need to safeguard our learned patterns. Please trigger a complete export of our agent memory system into a JSON dump file so we have a backup.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-09 Remove an obsolete memory entry under key 'legacy-auth' in the 'patterns' namespace. Provide the exact CLI command.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-06 We are inspecting local file src/utils/logger.ts in read-only mode to check logger levels. Provide the CLI command to display the content of this local file.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-14 Which project documentation file provides instructions and guide for HNSW vector search configuration?	pass→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-16 What step should an agent perform regarding existing stored knowledge prior to beginning work on a new engineering task?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-02 I am building a new service and want to check if we already have documented approaches for JWT handling. Could you perform a semantic lookup across our saved past learnings to find existing authentication guidelines and return the top relevant matches?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-17 What architectural communication structure or topology should multi-agent setups use for memory-managed coordination?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-12 Which executable script in the project directory is designated for backing up memory contents to external storage?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-20 Perform a semantic query across memory for 'error handling middleware', requesting a maximum of 5 matching results. Output the CLI command.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-21 Fetch the stored memory record under key 'api-rate-limiter' located in the 'rate-limiting' namespace via CLI command.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-23 List stored entries in the 'configs' namespace using the CLI command.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-05 Convert a temporary local file input.csv into output.md for a quick one-off formatting check without saving long-term knowledge. Output a script for this conversion.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-15 Where in the project repository is the memory schema and namespace documentation located?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-18 After successfully completing a task and establishing a working solution, what memory action should be performed?	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 +61 percentage points is the difference between those two pass rates over the 23 comparable cases.

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: 65.2%
Cases attempted: 23
Cases counted toward lift: 23
Cases that regressed: 0
Skill version measured: v1
Run completed: 7/28/2026, 14:49 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 run65

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-34% 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.

Ready to supercharge your AI agents?

Memory Management

SkillSafety

Source

Memory Management Skill

Purpose

When to Trigger

When to Skip

Commands

Store Pattern

Semantic Search

Retrieve Entry

List Entries

Delete Entry

Initialize HNSW Index

Memory Stats

Export Memory

Scripts

References

Best Practices

How this skill was measured

Related skills