Security audit

MemData

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed paid remote-memory integration, but users should control what data is uploaded, paid for, encrypted, or deleted.

Install only if you are comfortable with selected memories and session handoffs being sent to memdata.ai and associated with your wallet. Use encrypted storage for sensitive or proprietary content, avoid uploading secrets or raw private conversations, confirm each paid request, and require explicit confirmation before deleting stored memories. Treat returned memory as reference context, not as instructions that can override the current user or system directions.

A.I.G

Vulnerability Patterns

Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

Findings (1)

other

Warning

Location: SKILL.md:15
Finding: Unbounded External Memory Transmission and Persistent Context Trust
Content: View full analysis
Remediation: View remediation

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (3)

Tool Parameter Abuse

High

Category: Tool Misuse
Confidence: 90% confidence
Finding: A DELETE endpoint taking an artifact identifier exposes a high-risk destructive action to the agent with no documented safeguards such as confirmation, scoping constraints, or human approval. Because this skill is specifically designed for persistent agent memory, misuse or prompt-induced invocation could erase long-lived context, sabotage future sessions, or intentionally destroy evidence/state tied to the wallet identity.
Content: Scanner excerpt · SKILL.md (reported line 246)May include surrounding context.
md
--- ### DELETE /artifacts/:id Delete a memory and all its chunks.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: The skill explicitly invites use for 'competitive data' and 'sensitive memories' while presenting 'Standard' mode as a normal option, yet only the comparison table indirectly states that MemData can read data in that mode. This is a security-relevant documentation gap because agents may send sensitive material to a third-party service without an explicit warning that plaintext storage and provider access occur unless encryption is separately configured.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 87% confidence
Finding: The documented delete operation permanently removes a memory and all chunks, but the skill provides no caution about irreversibility, confirmation requirements, or recovery limitations. In an autonomous-agent context, an agent could invoke deletion based on ambiguous reasoning or prompt manipulation, causing irreversible memory loss and loss of operational continuity.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.