ClawHub
Back to skill

Security audit

Pilates (Session Planner, Form Coach, Progress Tracker)

Security checks for vulnerabilities and agentic risk

Overview

This is a Markdown-only Pilates planning skill that stores optional local practice notes with disclosed confirmation requirements and no external data sharing.

Before installing, be aware that this skill can create local notes about exercise tolerance, pain, pregnancy or postpartum context, surgery, and other safety limits under ~/pilates/. Use it only if you are comfortable storing that information locally, and do not rely on it as medical or rehabilitation care.

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

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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
Findings (4)

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 104)May include surrounding context.

md
## Common Traps

- Treating Pilates like a burn-more workout instead of a control practice -> speed replaces precision.
- Using advanced exercise names without checking the user's experience -> confusion and poor adherence.
- Correcting breathing, ribs, pelvis, neck, and pace all at once -> overload and lower body awareness.
- Copying reformer intensity onto mat work without context -> poor exercise selection.
- Forcing neutral spine as a rigid rule in every drill -> unnecessary tension and weaker movement quality.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill explicitly creates persistent files for health-adjacent information, including pain notes, pregnancy or postpartum notes, bone health or surgery notes, symptoms, and safety modifications, but provides no warning, consent step, minimization guidance, or privacy controls for storing this sensitive data on disk. This can lead to unintended retention or exposure of personal health information in a local filesystem, backups, sync services, or shared environments.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

This guidance directs exercise selection, progression, and symptom-based mode switching, including a repair mode for exercises that feel wrong or irritating, but it does not tell users when to stop and seek qualified medical advice for pain, injury, or health conditions. In a fitness and body-mechanics context, that omission can encourage users to self-manage potentially significant musculoskeletal issues or exercise through unsafe symptoms.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill instructs the agent to activate on a wide set of loosely related cues such as Pilates, mat work, core control, and posture-focused routines. This can cause unintended activation in unrelated or sensitive health conversations, leading the agent to steer interactions toward exercise guidance when it is not appropriate.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.