v1.0.0

Andrew Memory Layer

ReviewClawScan verdict for this skill. Analyzed May 1, 2026, 8:29 AM.

Analysis

The skill matches its memory-layer purpose, but it defaults to sending memory text to MiniMax and persists distilled conversation memories across sessions without clear review or deletion controls.

GuidanceReview this carefully before installing. If your conversations or memories may contain private information, prefer llmMode "local" with Ollama, avoid storing secrets, and make sure you have a way to inspect and delete saved memories. If using MiniMax, understand that memory text and search queries may be sent to that provider under your API key.

Findings (3)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

Permission boundary

Checks whether tool use, credentials, dependencies, identity, account access, or inter-agent boundaries are broader than the stated purpose.

Identity and Privilege Abuse

SeverityMediumConfidenceHighStatusNote

src/MemoryClient.js

this.apiKey = process.env.MINIMAX_API_KEY || ''; ... 'Authorization': `Bearer ${this.apiKey}`

The code uses an environment-provided MiniMax API key to authenticate provider calls; this is expected for cloud mode but is not reflected in the registry's declared required env vars or primary credential.

User impactInstalling users may not realize the skill can use their MiniMax account/API key and send requests under that credential.

RecommendationDeclare MINIMAX_API_KEY as an optional credential in metadata and explain when it is used.

Sensitive data protection

Checks for exposed credentials, poisoned memory or context, unclear communication boundaries, or sensitive data that could leave the user's control.

Insecure Inter-Agent Communication

SeverityHighConfidenceHighStatusConcern

src/MemoryClient.js

this.llmMode = options.llmMode || 'api'; ... fetch('https://api.minimaxi.com/v1/embeddings', ... 'Authorization': `Bearer ${this.apiKey}` ... body: JSON.stringify({ model: 'embo-01', texts: [text], type: 'query' }))

API mode is the default and sends the supplied memory or search text to MiniMax for embeddings, crossing a provider boundary with potentially sensitive long-term memory content.

User impactPersonal facts, identity text, or search queries stored in memory may be sent to MiniMax when the tool is used.

RecommendationUse local mode for private data, make cloud transmission an explicit opt-in, and clearly document what text is sent to MiniMax.

Memory and Context Poisoning

SeverityMediumConfidenceHighStatusConcern

src/MemoryClient.js

const conversation = messages.map(m => `${m.role}: ${m.content}`).join('\n'); ... const response = await this._callLLM(prompt); ... await this.add({ text: cleaned, memoryType: 'distilled', ... sourceFile: 'distill' });

Conversation content is summarized by an LLM and then stored as persistent memory without an artifact-shown approval or review step, so misleading or injected conversation text can become future context.

User impactBad or manipulative conversation content could be converted into long-term memories that influence later agent behavior.

RecommendationRequire user confirmation before saving distilled memories, keep provenance visible, and provide review/delete controls for stored memories.