substack-ghostwriting

Pass

Audited by Gen Agent Trust Hub on Apr 30, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted external data, which creates a surface for indirect prompt injection where embedded instructions could potentially influence agent behavior.\n
  • Ingestion points: Content is fetched from user-provided URLs using the WebFetch tool in Phase 0 of the Writing Workflow. Additionally, the references/voice-matching.md file directs the agent to analyze transcripts, social media posts, and Slack messages for voice extraction.\n
  • Boundary markers: There are no instructions specifying the use of delimiters or directives to ignore instructions contained within the analyzed source material.\n
  • Capability inventory: The skill leverages several powerful tools, including WebFetch, WebSearch, Write, Edit, and Agent (which allows the agent to call other skills, such as the mentioned humanizer skills).\n
  • Sanitization: No sanitization, filtering, or validation of the fetched or provided text content is performed before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 30, 2026, 02:33 AM