Code Pluginsource linked

MCP Memory Servicev0.1.5

Native OpenClaw plugin for mcp-memory-service: persistent memory with hybrid search, knowledge graph, and mistake notes

@filhocf/openclaw-mcp-memory·runtime mcp-memory·by @filhocf
Community code plugin. Review compatibility and verification before install.
openclaw plugins install clawhub:@filhocf/openclaw-mcp-memory
Latest release: v0.1.5Download zip

Capabilities

configSchema
Yes
Executes code
Yes
HTTP routes
0
Plugin kind
memory
Runtime ID
mcp-memory

Compatibility

Built With Open Claw Version
2026.5.20
Min Gateway Version
2026.3.24-beta.2
Plugin Api Range
>=2026.3.24-beta.2
Plugin Sdk Version
2026.5.20
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Suspicious
high confidence
!
Purpose & Capability
The persistent memory, search, storage, and prompt recall features match the stated purpose, but default auto-capture plus captureMatcher "*" makes the collection boundary broader than a typical user may expect.
!
Instruction Scope
The artifacts disclose auto-capture and auto-recall, but they do not provide clear consent, review, filtering, per-tool allowlisting, or sensitivity controls before storing tool outputs and injecting memories into later prompts.
Install Mechanism
Installation is a normal npm/OpenClaw plugin install with native dependencies such as better-sqlite3 and no install-script or shell persistence found in the inspected package.
!
Credentials
Local SQLite storage and local embedding are coherent for a memory plugin, but broad automatic ingestion of arbitrary tool results is disproportionate without tighter scoping because tool outputs can include secrets, files, account data, or private work product.
!
Persistence & Privilege
The plugin persistently stores memory entries, recalls them into future prompts, logs snippets of captured content and prompt queries, and offers irreversible deletion by hash without a confirmation flow.
Scan Findings in Context
[SQP-2] expected: Persistent storage and recall are expected for a memory plugin, but the scanner's concern is supported because autoCapture and autoRecall default to true and the artifacts lack clear privacy, retention, review, and deletion guidance.
[SQP-1] unexpected: The wildcard captureMatcher default of "*" is broader than necessary for the stated purpose and supports a Review verdict because it can capture unrelated tool outputs by default.
[SSD-3] unexpected: The broad heuristic extraction of strings and common output fields from arbitrary tool results is not well bounded and can persist sensitive fragments without semantic filtering.
[SQP-2] expected: The embedding-related concerns are partly mitigated by the artifacts describing a local ONNX model via @xenova/transformers, but prompt and tool-output text is still processed automatically and should be disclosed and controllable.
What to consider before installing
Install only if you want an always-on memory layer. Before enabling it, set autoCapture to false or restrict captureMatcher to a narrow allowlist, avoid using it around secrets or regulated data, and periodically review or delete stored memories. This review does not find evidence of malware or intentional exfiltration, but the default data-retention behavior is too broad to approve without user review.

Verification

Tier
source linked
Scope
artifact only
Summary
Validated package structure and linked the release to source metadata.
Commit
0b2298bc4fca
Tag
v0.1.5
Provenance
No
Scan status
suspicious

Tags

latest
0.1.5

openclaw-mcp-memory

Native OpenClaw plugin for mcp-memory-service: persistent memory with hybrid search, knowledge graph, and mistake notes.

Status

MVP em desenvolvimento.

Estrutura

src/
├── index.ts              # Entry point (definePluginEntry)
├── hooks/
│   ├── auto-recall.ts    # before_prompt_build → injetar memórias
│   ├── auto-capture.ts   # after_tool_call → extrair fatos
│   └── session-end.ts    # stop → consolidar
├── tools/
│   ├── memory-store.ts   # Armazenar memória
│   ├── memory-search.ts  # Busca híbrida
│   ├── memory-forget.ts  # Remover memória
│   └── memory-stats.ts   # Estatísticas
└── storage/
    ├── sqlite.ts         # Adaptador SQLite
    ├── embeddings.ts     # Embeddings ONNX
    └── graph.ts          # Knowledge Graph

Config

{
  plugins: {
    entries: {
      "mcp-memory": {
        enabled: true,
        config: {
          model: "local",
          threshold: 0.7,
          maxResults: 5,
          autoCapture: true,
          autoRecall: true,
        },
      },
    },
  },
}