auth-flow-operator
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [NO_CODE]: The skill is entirely composed of markdown documentation and instructional text. No executable scripts, binaries, or automated shell commands are included in the package.- [INDIRECT_PROMPT_INJECTION]: The skill provides a surface for indirect prompt injection due to its core function of interacting with external target URLs, which is a necessary component of its intended use case.
- Ingestion points: The
target_urlandauth_notesinputs defined in SKILL.md. - Boundary markers: No explicit delimiters or instructions to ignore embedded content are specified.
- Capability inventory: The workflow involves route discovery and manual authentication validation.
- Sanitization: No sanitization logic is provided for external content encountered during testing.- [SAFE]: The instructions include proactive security measures, such as explicit constraints against brute-force attacks and requirements to minimize the logging of PII and credentials.
Audit Metadata