functional-area-resolver
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to run local verification tools such as
gbrain routing-evaland Node.js scripts likeharness.mjs. These are used to ensure routing accuracy after compression. - [INDIRECT_PROMPT_INJECTION]: The skill processes content from local routing files, creating a surface for potential injection if those files contain untrusted data.
- Ingestion points: The skill reads
RESOLVER.mdandAGENTS.mdfiles (SKILL.md). - Boundary markers: Absent; the skill does not use specific delimiters to isolate the content of the routing files from its own instructions.
- Capability inventory: The skill utilizes
exec,read,write, andedittools to modify files and execute verification commands (SKILL.md). - Sanitization: Absent; the skill performs structural transformations but does not explicitly sanitize the input strings for embedded instructions.
Audit Metadata