compose
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The documentation provides instructions to install the skill using
npxor by downloading a.skillfile from the author's GitHub repository (github.com/synapseradio). These are standard methods for skill distribution and point to the vendor's own infrastructure. - [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided subjects and reference data. This presents an attack surface for instructions embedded in external data.
- Ingestion points: The user-supplied
subjectand the facts provided in theGivensection of the prompt. - Boundary markers: The skill uses markdown headers and bold labels to structure output, but does not explicitly instruct the agent to ignore instructions embedded in the inputs.
- Capability inventory: The skill has no access to external tools, file system operations, or network capabilities.
- Sanitization: No explicit sanitization of input data is performed.
- Impact: Given the lack of tool access, the impact of an injection is limited to affecting the reasoning output within the conversation.
Audit Metadata