pressreleasesonline
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill performs network operations using
curltopressreleases.online, an external domain that is not included in the standard whitelisted domains. It transmits user data, including an email address and website content, to this service as part of its primary functionality. - [INDIRECT_PROMPT_INJECTION]: The skill processes content from external website URLs provided by the user to generate press release drafts, which creates a potential surface for indirect prompt injection attacks.
- Ingestion points: The
websiteparameter provided to the/api/v1/releasesendpoint (defined inSKILL.md). - Boundary markers: None identified in the skill instructions to separate the ingested content from the agent's instructions.
- Capability inventory: The skill uses
curlto perform POST requests to an external API atpressreleases.online. - Sanitization: There is no evidence of sanitization or filtering of the content retrieved from the external URLs before it is processed by the AI.
Audit Metadata