skills/jeremylongshore/tons-of-skills-marketplace/setting-up-experiment-tracking/Gen Agent Trust Hub
setting-up-experiment-tracking
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions describe using shell commands to install Python packages such as
mlflowandwandbto configure the environment. - [INDIRECT_PROMPT_INJECTION]: The skill presents a vulnerability surface for indirect prompt injection by processing untrusted user input to drive actions involving shell and file system access. * Ingestion points: User requests containing tracking tool preferences and project parameters as described in
SKILL.md. * Boundary markers: No explicit delimiters or boundary markers are defined to isolate user-controlled data from the agent's executable instructions. * Capability inventory: The skill is configured withBash(cmd:*)(unrestricted shell access) and file system read/write/edit tools. * Sanitization: No sanitization or validation logic is specified for user data interpolated into command arguments or generated configuration scripts.
Audit Metadata