Archive Project

• Outcome: Pass
• Latest audit: May 28, 2026, 4:41 AM
• Version:
  1.2.5

Overview

This skill is a disclosed project-archiving workflow, but it handles sensitive local transcripts and should be used only when the user intends long-term retention.

Install only if you want completed-project transcripts retained in a local archive. Before committing, confirm the selected transcript, review the sanitized output, avoid storing unnecessary client contact details, and approve deletion of original session files only after verifying the archive copy is correct.

SkillSpector

By NVIDIA

Vulnerability Patterns

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

Findings (5)

Scope Creep

High

• Confidence: 96% confidence
• Finding: The skill documents deletion of session transcripts from the configured session store even though only read access is declared for that path. This creates a dangerous permission/behavior mismatch: an operator may approve or trust a workflow that can destroy audit history or user data without the manifest accurately signaling that destructive capability.

Vague Triggers

Medium

• Confidence: 88% confidence
• Finding: The trigger phrases "archive this" and "can we archive this" are broad enough to match ordinary conversation and may cause the skill to run when the user did not explicitly intend archival. In this skill's context, unintended execution is meaningful because it reads session transcripts, writes archive files, and performs git commits, so accidental activation can cause privacy and integrity issues even without malicious intent.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The README states that the skill creates archive directories, updates files, and git-commits to the workspace, but it does not prominently warn users that running the skill will make persistent modifications. Because the skill also ingests session transcripts from a configurable path, insufficient disclosure increases the risk of users triggering archival without realizing sensitive data will be copied into long-term storage and committed.

Vague Triggers

Medium

• Confidence: 81% confidence
• Finding: The trigger guidance includes broad conversational cues such as deliveries, topic changes, or task transitions that can cause unsolicited archive prompts. In a skill that reads and persists session transcripts, premature prompting increases the chance of archiving sensitive or incomplete work without clear user intent, creating privacy and workflow integrity risks.

Natural-Language Policy Violations

Medium

• Confidence: 94% confidence
• Finding: The archive template instructs recording client contact information for future reference without requiring minimization, opt-in, or policy justification. Persisting personally identifiable information in project archives broadens data retention and increases harm if the workspace is later accessed, synced, or exposed.

VirusTotal

VirusTotal findings are pending for this skill version.