runpod-cloud-migrate
Pass
Audited by Gen Agent Trust Hub on Nov 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the
Bashtool to execute a local Python script (scripts/scan_cloud_usage.py) and therg(ripgrep) utility. These tools are used solely for scanning the user's local repository to identify cloud service usage signals (SDKs, CLIs, configurations). This behavior is constrained to the skill's intended purpose and the user-provided workspace. - [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it processes untrusted data from the user's repository.
- Ingestion points: The
scripts/scan_cloud_usage.pyscript walks the target directory and reads the contents of all detected text files to match them against regex patterns. - Boundary markers: The scanner output is structured as Markdown or JSON for the agent to interpret, providing some separation from the raw file content.
- Capability inventory: The agent has the capability to interface with the Runpod API and various cloud CLIs to provision or terminate resources.
- Sanitization: While the scanner extracts matching lines which could theoretically contain malicious instructions, the skill's global rules (specifically Rules 2, 3, and 8 in
SKILL.md) mandate explicit human-in-the-loop approval for all billable, destructive, or critical migration steps, effectively mitigating the risk of automated exploitation.
Audit Metadata