onboarding-cro
Pass
Audited by Gen Agent Trust Hub on May 7, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [SAFE]: Analysis of the skill instructions, evaluation criteria, and reference documentation found no evidence of malicious patterns, obfuscation, or unauthorized command execution.
- [DATA_EXFILTRATION]: The skill specifies that the agent should read from
.agents/product-marketing-context.mdor.claude/product-marketing-context.mdto gain product context. This behavior is localized to the agent's environment and is not associated with any network capabilities or credentials that would enable data exfiltration. - [PROMPT_INJECTION]: The skill utilizes data from external context files, creating a theoretical surface for indirect prompt injection.
- Ingestion points:
.agents/product-marketing-context.mdand.claude/product-marketing-context.md(referenced in SKILL.md). - Boundary markers: None identified; the skill instructs the agent to read the files directly.
- Capability inventory: The skill has no active capabilities, tools, or scripts (no file-writing, no network access, no subprocess execution).
- Sanitization: None; the content is read and used directly to inform the agent's recommendations.
Audit Metadata