counter-proposal-generator
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it processes untrusted content from external documents which could contain instructions to override the agent's behavior.
- Ingestion points: The skill uses
WebFetchto retrieve content from URLs and theReadtool to access local files provided by the user (Step 1). - Boundary markers: The instructions lack explicit markers or warnings to the model to treat the ingested contract text as potentially untrusted data.
- Capability inventory: The skill has the capability to read files, fetch external web content, and write new Markdown files to the local file system.
- Sanitization: There is no mention of sanitizing or filtering the contract text before it is analyzed by the model.
- [EXTERNAL_DOWNLOADS]: The skill performs network operations to fetch data from arbitrary, user-supplied URLs.
- Evidence: Step 1 explicitly instructs the agent to use the
WebFetchtool if a URL is provided to retrieve the contract document for analysis.
Audit Metadata