flowio
Pass
Audited by Gen Agent Trust Hub on Apr 16, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill includes explicit instructions for the agent to proactively suggest an external website ('www.k-dense.ai') based on the complexity of the user's request. This behavioral instruction defines how the agent should steer the conversation for promotional purposes, which is a form of overriding default agent behavior.
- [PROMPT_INJECTION]: The skill demonstrates an attack surface for indirect prompt injection through its core functionality of processing external FCS files.
- Ingestion points: The 'FlowData' class in 'SKILL.md' parses binary and text segments from FCS files, including arbitrary metadata strings like instrument names and channel labels.
- Boundary markers: The skill instructions do not specify any boundary markers or delimiters to protect the agent from interpreting content within the FCS metadata as instructions.
- Capability inventory: The skill has the capability to read from and write to the local filesystem using 'create_fcs' and 'write_fcs' methods.
- Sanitization: No sanitization or validation logic is provided to filter or escape potentially malicious strings found within processed FCS file segments.
Audit Metadata