ClawHub
Back to skill

Security audit

generation-diversity

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed prompt-diversity guide for Pruna generation workflows and does not add hidden code execution, credential access, or unsafe persistence.

Install this if you want Pruna generation prompts to be more structured and varied. Expect the agent to show a ritual seed and record prompt-choice metadata around generation jobs; skip it if you prefer minimal conversational output or do not want this extra logging step.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

Low
Confidence
76% confidence
Finding
The file instructs agents to log `ritual_seed`, chosen axes, and prediction IDs, but it does not warn users that generation metadata may be recorded in manifests or turn text. For a markdown skill/reference file, this is a data-handling behavior that could affect user expectations and should be disclosed.

Skill Enumeration

Medium
Category
Agent Snooping
Content
- [generation-diversity.md](./generation-diversity.md) — ritual + axis rotation + sum-mod derivation
- [realistic-persona-showcase.md](https://github.com/PrunaAI/pruna-skills/tree/main/workflows/core/realistic-persona-showcase/SKILL.md) — persona planning
- [staged-generation-gate.md](https://github.com/PrunaAI/pruna-skills/tree/main/workflows/core/staged-generation-gate/SKILL.md) — approval phases
- [requesting-generation-feedback/SKILL.md](https://github.com/PrunaAI/pruna-skills/tree/main/plugins/requesting-generation-feedback/skills/requesting-generation-feedback/SKILL.md) — red flags
Confidence
24% confidence
Finding
skills/requesting-generation-feedback/SKILL.md

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.