ClawHub
Back to skill

Security audit

team-memory

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed local team-memory tool for sensitive personnel notes, with no evidence of hidden network access, credential use, destructive behavior, or background persistence.

Install only if you are comfortable storing sensitive team, performance, 1:1, and stakeholder feedback locally. Keep the `data/` folder and exports out of public repositories or shared drives unless properly protected, review dashboard outputs before sharing, and use the dry-run/confirmation flows for imports, migrations, and task resolution.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Lp3

Medium
Category
MCP Least Privilege
Confidence
94% confidence
Finding
The skill instructs the agent to read and write local files and execute shell/Python scripts, but it does not declare permissions or present clear capability boundaries. Because the data handled includes sensitive HR-style notes, stakeholder complaints, and performance evidence, undeclared file and shell access materially increases the risk of unauthorized modification, disclosure, or destructive actions if the skill is invoked in the wrong context or misused.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The template explicitly instructs storing real names in record bodies and in `skill-config.yaml`, which increases exposure of personally identifiable information across multiple files and metadata locations. In a team-memory skill that stores performance notes, 1:1 conversations, complaints, and stakeholder feedback, this creates unnecessary privacy risk, broadens the blast radius of accidental disclosure, and may conflict with data-minimization expectations even if the system is local-only.

Vague Triggers

Medium
Confidence
92% confidence
Finding
The trigger phrases are very broad everyday managerial terms such as '团队看板' and '一张图了解团队', so this reference may activate in situations far beyond the author's intended scope. Over-broad invocation can cause the agent to run local scripts and assemble sensitive team-memory data in response to generic requests, increasing the chance of unintended disclosure or inappropriate use of personnel-related information.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

Detected: suspicious.dynamic_code_execution

Dynamic code execution detected.

Critical
Code
suspicious.dynamic_code_execution
Location
scripts/team-dashboard.py:31