Dont Shoot The Dog The New Art Of Teaching And Training

• Outcome: Review
• Latest audit: Jun 13, 2026 at 1:24 AM UTC
• Version:
  1.0.1

Overview

This is a text-only behavior-training skill, but it needs review because it gives broad advice for shaping human behavior without clear consent or safety guardrails.

Review this skill before installing if you may use it for parenting, workplace, relationship, or other human behavior advice. It is technically low-risk, but users should apply it only transparently, with consent and respect for autonomy, and should avoid physical interventions or covert behavior-shaping.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (7)

Vague Triggers

Medium

• Confidence: 96% confidence
• Finding: The recall triggers are extremely broad and map to common, everyday situations like parenting, communication, habit change, and frustration with behavior problems. This increases the chance the skill is invoked outside its intended scope, causing unsolicited behavior-modification guidance to surface in sensitive interpersonal contexts.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The skill repeatedly frames behavior-shaping techniques as applicable to children, colleagues, family members, and relationships without caution about consent, autonomy, or manipulation risks. In this context, the content could normalize covert influence or emotional conditioning in human relationships, especially when presented as a general-purpose method for changing others.

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: This section explicitly presents coercive and potentially harmful behavior-change methods, including punishment, negative reinforcement, and 'Shoot the Animal,' without any safety framing, ethical caution, or guidance on misuse. In a general-purpose agent skill, such content can normalize manipulative or abusive application to humans or animals and may be repurposed to justify harmful interventions.

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: The document provides detailed behavior-modification techniques that can be applied to humans and animals, including extinction, variable reinforcement, and controlling unwanted behavior, without any safety, consent, or ethical-use guidance. In a general-purpose agent skill, this can enable manipulative or coercive interpersonal use, especially because the text frames the methods as broadly applicable to children, employees, and other people.

Missing User Warnings

Medium

• Confidence: 87% confidence
• Finding: The content presents a training exercise involving real people and includes a case where one participant physically trips another to force a training outcome, but it provides no safety warning, boundaries, or prohibition on unsafe physical contact. In a practical how-to document, readers may imitate the example and normalize risky behavior, creating a foreseeable risk of falls or injury.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: This section explicitly advises covertly shaping another person's behavior ('shape silently' and 'never announce your training plan') without any consent, autonomy, or manipulation warning. In an agent skill or instructional context, that normalizes deceptive influence techniques that can be applied to manipulate roommates, coworkers, or users in ways that bypass informed consent.

Missing User Warnings

Medium

• Confidence: 92% confidence
• Finding: This passage presents successful training of a person without their awareness as an illustrative lesson, but omits any ethical framing about consent, dignity, or potential abuse. That makes the content more dangerous because it can be read as endorsement of nonconsensual psychological conditioning rather than historical or critical discussion.

VirusTotal

VirusTotal findings are pending for this skill version.