Tinmem Memory System

Security checks across static analysis, malware telemetry, and agentic risk

Overview

The skill is clearly meant for memory, but it automatically stores and reuses personal information across sessions without clear limits, consent, or retention controls.

This appears to be a legitimate memory skill, but it should be reviewed carefully because it stores user-related information persistently and automatically reuses it. Use it only if you are comfortable with cross-session memory, and verify that you can inspect, correct, delete, and limit what it remembers.

Static analysis

No static analysis findings were reported for this release.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Risk analysis

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.

What this means

Personal or sensitive details shared in one conversation could be saved and later used in other conversations, including if the saved memory is outdated, incorrect, or contains harmful instructions.

Why it was flagged

The skill documents automatic cross-session storage and reuse of memory, but does not specify consent, exclusions, retention, review, or safeguards against bad or sensitive memories influencing future responses.

Skill content
Memories are automatically injected into context before each response via `<agent-experience>` tags
- New memories are automatically extracted after each conversation turn
- All memories persist across sessions in a local LanceDB database
Recommendation

Install only if you want automatic persistent memory; before use, look for controls to review, edit, delete, disable, or limit what is stored and when memories are injected.

What this means

The assistant may retain sensitive personal attributes longer or more broadly than the user expects.

Why it was flagged

The memory categories explicitly include identity and demographic information, and the artifact does not describe restrictions for sensitive personal data or user approval before saving it.

Skill content
`profile` | User identity, role, expertise, demographics | Always merge
Recommendation

Avoid sharing sensitive personal information unless the memory system provides visible consent, deletion, and exclusion controls.