ClawHub
Back to skill

Security audit

Test B

Security checks for vulnerabilities and agentic risk

Overview

This instruction-only skill has a coherent purpose, but it may persist sensitive conversation and tool-output details too broadly without clear user approval or redaction rules.

Install only if you are comfortable with the agent scanning the active conversation and tool outputs and appending selected details to persistent memory files. Before using it on sensitive work, add or require rules that exclude secrets, tokens, credentials, cookies, private keys, personal data, and sensitive configuration values, and require explicit confirmation before any memory write.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The README explicitly promotes extracting important conversation details and writing them into persistent memory files before compaction, but it does not warn that those details may include secrets, personal data, internal endpoints, or other sensitive operational context. In an agent workflow, this can increase unintended data retention and broaden exposure if memory files are later accessed, synced, or reused across tasks.

Vague Triggers

Medium
Confidence
85% confidence
Finding
The skill defines broad and partially automatic activation conditions such as triggering when context is near full or when a heartbeat detects 80% usage. That can cause the skill to run unexpectedly in unrelated conversations, increasing the chance of unintended memory writes, user confusion, and premature compaction workflows without clear user intent.

Ssd 3

Medium
Confidence
95% confidence
Finding
The skill instructs the agent to extract broadly scoped facts, configurations, endpoints, file paths, errors, and user preferences from tool output and append them into persistent memory files. Because there is no filtering for secrets, credentials, tokens, personal data, or environment-sensitive details, the compaction workflow can transform transient sensitive data into durable storage, increasing the blast radius and likelihood of later disclosure.

Static analysis

No suspicious patterns detected.