free-tool-strategy
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to incorporate external context from a file named
marketing-context.mdif it exists in the workspace. This creates a surface where instructions embedded in the context file could potentially influence the agent's behavior. - Ingestion points:
SKILL.mdcontains instructions for the agent to readmarketing-context.mdto gather background on products, audience, and goals. - Boundary markers: The skill does not define specific delimiters or instructional warnings to ignore potential commands embedded within the context file.
- Capability inventory: The skill provides a Python script (
scripts/tool_roi_estimator.py) that performs local mathematical calculations. It does not include tools for network communication, file system modification, or privilege escalation. - Sanitization: There is no evidence of content filtering or validation performed on the ingested context data before it is processed by the agent.
Audit Metadata