coss-particles
Pass
Audited by Gen Agent Trust Hub on Oct 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to fetch and process external data from remote URLs, which constitutes a potential attack surface for indirect prompt injection.
- Ingestion points: The skill provides hundreds of links to remote JSON manifests (e.g.,
https://coss.com/ui/r/p-accordion-1.json) intended to be fetched and processed. - Boundary markers: There are no instructions for the agent to use boundary markers or to ignore instructions that might be embedded within the retrieved JSON data.
- Capability inventory: The agent is prompted to fetch, read, and 'adapt' source code from these external manifests, which involves interpreting and reproducing code in the project environment.
- Sanitization: The skill does not include any guidance for sanitizing, validating, or filtering the content of the external JSON files before the agent processes it.
Audit Metadata