ClawHub
Back to skill

Security audit

Daniels Running Plan

Security checks for vulnerabilities and agentic risk

Overview

This is an instruction-only running-plan helper whose local file export is disclosed and purpose-aligned, though users should notice that generated plans may be saved on disk.

Before installing, be aware that generated training plans may be saved as TXT and CSV files on the desktop path named by the skill. Use chat-only output or request a different filename/path if you do not want local files, and avoid sharing fitness or schedule details you would not want stored locally.

SkillSpector

Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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)

Description-Behavior Mismatch

Medium
Confidence
94% confidence
Finding
The skill’s stated purpose is to generate a running training plan, but it also instructs writing files to a fixed desktop path. This introduces unnecessary side effects and local persistence that are not essential to answering the user, creating privacy and security risk if personal fitness data is saved without explicit consent.

Context-Inappropriate Capability

Medium
Confidence
96% confidence
Finding
Granting or encouraging local filesystem writes is not justified by a running-plan skill whose core function is text generation. Unnecessary write capability expands the attack surface, enables unintended data persistence, and could be abused by a compromised or repurposed skill to drop files in user-accessible locations.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The documentation instructs saving output to desktop files without telling the user that data will be written locally. Because the collected inputs include age, sex, schedule, and performance data, silent persistence can violate user expectations and create avoidable privacy exposure on a shared or monitored device.

Static analysis

No suspicious patterns detected.