Back to skill

Security audit

TranscriptLab

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rated the skill benign with high confidence; scanners found one medium-severity privacy disclosure concern in an otherwise complete scan of the Markdown skill.

Consider whether you are comfortable sending the media you submit to TranscriptLab for processing. The scan covered the skill’s single Markdown component and reported complete coverage; benign does not mean risk-free.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • 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)

The file-transcription workflow lacks a warning that attached media is sent to an external service.

Medium
Category
Output Handling
Confidence
90% confidence
Finding

The routing instructions direct the agent to submit attached files or media URLs to TranscriptLab for transcription without warning that potentially private media will be shared with that service. The statement that results enter the user's TranscriptLab library discloses output storage, but not the transfer of source media; the instruction to avoid discussing the plumbing further discourages that disclosure.

Content

No source excerpt is available for this finding.