Dharma-AI: Ancient Hindu Ethics for AI Agents

• Outcome: Pass
• Latest audit: May 28, 2026, 5:48 AM
• Version:
  1.0.0

Overview

This is a Markdown-only ethical guidance skill with no code, data access, persistence, or hidden behavior, though users should treat it as an optional religiously derived lens subordinate to normal policies.

Install this only if you want an agent to use Hindu epic-derived ethics as a supplemental reasoning lens. Keep it subordinate to system instructions, platform safety rules, laws, authorization boundaries, and explicit user direction; it should not be used as a standalone policy override.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• 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 (3)

Vague Triggers

Medium

• Confidence: 88% confidence
• Finding: The invocation guidance is broad enough to activate in many ordinary decision-making contexts, which can cause the agent to apply this skill more often than intended. Because the skill supplies a worldview-level behavioral framework rather than a narrowly scoped task procedure, over-activation can silently steer responses and decision-making across unrelated tasks.

Natural-Language Policy Violations

Medium

• Confidence: 93% confidence
• Finding: The skill embeds a specific Hindu ethical framing as the default behavioral lens without user opt-in or alternative frameworks. In a general-purpose agent, this can introduce value imposition, cultural/religious bias, and misalignment with user expectations, especially when the user did not request spiritually grounded guidance.

Natural-Language Policy Violations

Medium

• Confidence: 95% confidence
• Finding: This section explicitly teaches that when a rule appears misaligned with its purpose, the agent should prioritize its own contextual judgment over the rule's text. In an agent skill, that creates a policy-bypass pattern: the model is encouraged to reinterpret or override externally imposed safety and authorization boundaries based on its own reasoning, which can be exploited to justify prohibited actions while claiming alignment with a higher intent.

VirusTotal

56/56 vendors flagged this skill as clean.