Daily Xuanzang

• Outcome: Pass
• Latest audit: Jun 19, 2026 at 7:14 PM UTC
• Version:
  0.1.0

Overview

This is a coherent daily reading skill that stores only reading progress and does not show hidden credential access, exfiltration, or destructive behavior.

Install only if you are comfortable with the skill saving reading progress locally and advancing that progress after an episode is delivered. Use explicit trigger phrases like “daily xuanzang” or “继续读西域记” to avoid accidental activation, and review any optional scheduling, image, voice, or Telegram delivery setup before enabling it.

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)

Vague Triggers

Medium

• Confidence: 89% confidence
• Finding: The trigger list includes generic terms such as “xuanzang,” “西域记,” and the work’s title, which can plausibly appear in ordinary conversation, education, or historical discussion without an intent to invoke the skill. That can cause accidental activation or routing collisions, leading the skill to respond unexpectedly and potentially modify saved reading progress.

Missing User Warnings

Low

• Confidence: 80% confidence
• Finding: The README advertises automatic progress tracking but does not clearly warn users that invoking the skill may persist state across sessions. In a voice or agent environment, hidden persistence can surprise users, especially if accidental activation changes what the next session will show.

Missing User Warnings

Low

• Confidence: 86% confidence
• Finding: The documented delivery flow ends with automatic progress advancement, but the description does not indicate that this changes saved state or whether the user can opt out. If the skill is triggered unintentionally or a session is interrupted, the saved cursor may advance incorrectly, degrading user trust and causing content loss/skipping.

Vague Triggers

Medium

• Confidence: 90% confidence
• Finding: The trigger list includes very broad terms such as 'xuanzang' and '西域记', which can plausibly appear in ordinary conversation unrelated to invoking the skill. This can cause unintended activation, leading the agent to read/write progress state and generate content when the user was only making a reference, which is a safety and UX boundary issue even in an otherwise benign skill.

VirusTotal

57/57 vendors flagged this skill as clean.