splitwise
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill uses
npxto download and execute the@useclawlink/clipackage from the NPM registry. This is a standard mechanism for utilizing the vendor's command-line tools to interact with the Splitwise API through their hosted service. - [INDIRECT_PROMPT_INJECTION]: The skill interacts with external data from Splitwise, creating a potential surface for indirect prompt injection if malicious data is present in expense descriptions or group names.
- Ingestion points: Data is ingested from the Splitwise API via the
describeandactionscommands inSKILL.md. - Boundary markers: The skill includes a clear instruction to "confirm with the user before any write," establishing a human-in-the-loop security boundary.
- Capability inventory: The skill possesses the capability to read and modify Splitwise data (e.g., creating expenses, settling bills) via the ClawLink CLI.
- Sanitization: The skill relies on the underlying CLI and the agent's internal filtering, as no specific sanitization logic is described for the retrieved data.
Audit Metadata