Praesidia

• Outcome: Review
• Latest audit: May 28, 2026 at 6:48 AM UTC
• Version:
  0.1.0

Overview

This Praesidia skill is coherent, but it needs review because it can change live agent guardrails and send validation text to Praesidia without clear confirmation or privacy prompts.

Install only if you intend your assistant to use your Praesidia account for agent verification and guardrail management. Use a scoped API key, confirm any guardrail changes before applying them to production agents, and avoid sending confidential or regulated content for validation unless sharing it with Praesidia is acceptable.

SkillSpector

By NVIDIA

Vulnerability Patterns

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

Findings (4)

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The README instructs the assistant to apply guardrails that can change blocking, redaction, replacement, or escalation behavior on a user's agent, but it does not require explicit user confirmation or warn that this is a configuration-changing action. In an agent skill context, this can cause unintended security or operational changes if the assistant treats a conversational request as authorization to modify production protections.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The content-validation flow tells the assistant to send arbitrary message content to Praesidia for analysis without warning that the submitted text leaves the local environment and may contain sensitive data. In a security skill, users may paste secrets, PII, regulated data, or proprietary prompts, so omission of a transmission/privacy notice materially increases disclosure risk.

Vague Triggers

Medium

• Confidence: 80% confidence
• Finding: The trigger text is broad enough to match generic security-related queries such as 'is this agent safe?' or 'apply guardrails,' which can cause the skill to activate outside narrow Praesidia-specific contexts. Overbroad routing can send unrelated user data or requests into this integration unnecessarily, increasing the chance of unintended external calls, confusion, or unauthorized actions against agent configurations.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The skill instructs the agent to perform a POST that changes guardrail configuration immediately, without requiring a confirmation step or warning the user that this modifies a live agent's security policy. That creates a real risk of unintended configuration changes, policy drift, or accidental disruption if the assistant misidentifies the target agent or the user's intent.

VirusTotal

66/66 vendors flagged this skill as clean.