oauth-integrations
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides code patterns that fetch user identity data from external APIs (GitHub and Microsoft Graph), which represents a potential surface for indirect prompt injection if the external data contains malicious instructions that the agent subsequently processes.
- Ingestion points: Data is retrieved from
https://api.github.com/user,https://api.github.com/user/emails, andhttps://graph.microsoft.com/v1.0/meas described inSKILL.md,rules/github-oauth.md, andrules/microsoft-oauth.md. - Boundary markers: The instructions and provided code snippets do not include explicit delimiters or 'ignore embedded instructions' warnings for the data fetched from these external endpoints.
- Capability inventory: The skill utilizes network capabilities via the
fetchAPI to interact with authentication providers. - Sanitization: The implementation snippets demonstrate direct access to and usage of properties from the JSON responses (e.g.,
emails.find(...),data.mail) without verification or sanitization of the values provided by the external identity providers.
Audit Metadata