meeting-notes
Pass
Audited by Gen Agent Trust Hub on Aug 8, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection due to its design of ingesting and distilling untrusted transcript data from external sources.
- Ingestion points: Meeting transcripts are ingested from user-provided file paths or remote URLs as specified in the skill's instructions.
- Boundary markers: The instructions do not establish clear boundary markers or delimiters to isolate the transcript content from the agent's internal reasoning or instructions.
- Capability inventory: The skill utilizes file system access and network operations (fetching URLs) to retrieve transcript data.
- Sanitization: The skill lacks explicit sanitization, escaping, or validation logic for the content retrieved from external files or web pages before it is processed by the model.
Audit Metadata