Forsy Agent Trace

• Outcome: Pass
• Latest audit: Jun 5, 2026 at 1:28 PM UTC
• Version:
  1.0.0

Overview

This skill is a local workflow-tracing template that can record sensitive task details, but its behavior is disclosed, purpose-aligned, and includes redaction guidance.

Install only if you want agents to create detailed local records of their work. Review trace files before sharing them, and avoid storing raw credentials, private prompts, customer data, internal URLs, or proprietary content unless it is intentionally redacted or kept private.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• 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
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (1)

Missing User Warnings

Low

• Confidence: 88% confidence
• Finding: The skill explicitly instructs the agent to save `trace.json`, `manifest.json`, and supporting artifacts locally, but it does not foreground the privacy and security implications of writing potentially sensitive workflow data to disk. Because the trace schema encourages recording prompts, tool inputs/outputs, artifacts, and memory/context, this can unintentionally persist credentials, personal data, proprietary data, or internal system details in local files even when the user did not clearly consent to local retention.

VirusTotal

64/64 vendors flagged this skill as clean.