NARRA Memory System

Security checks across malware telemetry and agentic risk

Overview

This is a markdown-only memory-organization skill that intentionally changes agent memory files, with no hidden execution, network use, or malware indicators found.

Install only if you want an agent to maintain a persistent narrative memory. Review edits to IDENTITY.md, MEMORY.md, AGENTS.md, and HEARTBEAT.md before relying on them, and avoid feeding sensitive logs or transcripts into the narrative unless you are comfortable preserving that information.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Memory PoisoningPersistent Context Injection, Context Window Stuffing, Memory Manipulation
  • 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
Findings (1)

Memory Manipulation

High
Category
Memory Poisoning
Content
- Include lessons learned and relationships formed
- End with current project status table

### Step 3: Rewrite IDENTITY.md

Expand from a simple name/role card to a complete identity anchor:
- Core identity (name, role, creature, vibe)
Confidence
80% confidence
Finding
Rewrite IDENTITY

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal