transcription-to-content
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external text content (transcripts) provided by the user, creating a surface where malicious instructions embedded in the data could influence agent behavior. However, the skill lacks tool-based capabilities (like network access or file system modifications) that would allow for successful exploitation.
- Ingestion points: User-supplied podcast transcripts processed via the 'Repurpose Episode Transcript', 'Create Show Notes', and 'Extract Social Content' prompts.
- Boundary markers: Absent. The templates instruct the user to 'paste transcript' directly into the prompt without utilizing delimiters or isolation instructions.
- Capability inventory: None. The skill provides workflow instructions and templates but does not execute subprocesses, perform file writes, or initiate network requests.
- Sanitization: None defined.
- [SAFE]: No malicious code, obfuscated payloads, or persistence mechanisms were detected. The skill uses standard markdown templates and refers to well-known productivity services like Notion and Airtable.
Audit Metadata