hallucination-guard
Pass
Audited by Gen Agent Trust Hub on Nov 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines a process requiring the agent to analyze local project files, such as
package.jsonand source code, to confirm the existence of libraries and methods. This constitutes a vulnerability surface where untrusted data enters the agent context.\n - Ingestion points: Specifically identifies
package.jsonand project source files as data sources to be analyzed.\n - Boundary markers: There are no instructions or delimiters provided to prevent the agent from being influenced by malicious instructions embedded within the comments or metadata of the files being verified.\n
- Capability inventory: The skill leverages the agent's file system search (
grep) and file reading (view_file) tools to perform its checks.\n - Sanitization: No content sanitization or verification logic is provided to distinguish between actual code signatures and adversarial prompts hidden in the codebase.
Audit Metadata