fraud-signal-prioritizer
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external business signals, which introduces a potential surface for indirect prompt injection if the input data contains instructions aimed at influencing the agent's behavior.
- Ingestion points: The
run.pyscript reads content from a file path provided through the--inputcommand-line argument. - Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are defined in the
SKILL.mdorrun.pyto isolate the data from the agent's instructional context. - Capability inventory: The skill performs local file system operations, including reading and writing files via the
pathlibmodule inrun.py. - Sanitization: There is no evidence of data sanitization or validation of the input JSON structure within the current implementation.
Audit Metadata