skills/aliyun/alibabacloud-aiops-skills/alibabacloud-migration-mas-cloud-migration-survey/Gen Agent Trust Hub
alibabacloud-migration-mas-cloud-migration-survey
Pass
Audited by Gen Agent Trust Hub on Jul 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from user-provided files which could contain hidden instructions targeting the LLM.
- Ingestion points:
scripts/extract.pyreads content from Excel (.xlsx), Word (.docx), CSV, and text files provided by the user. - Boundary markers: The skill does not implement specific delimiters or escape sequences to separate processed file content from system instructions when presented to the LLM.
- Capability inventory: The skill possesses the ability to execute subprocesses (calling its own Python scripts), read from the local filesystem, and write files to disk (
scripts/generate_report.pyandscripts/prepare_sample_input.py). - Sanitization: The skill includes strong path traversal protections (checking for
../and null bytes usingos.path.realpath) in all scripts. Additionally,scripts/generate_report.pyperforms a quality check that scans for sensitive data like AWS/Alibaba Cloud AccessKeys and plaintext passwords using regular expressions. - [COMMAND_EXECUTION]: The skill instructions mandate the execution of specific local Python scripts to perform its tasks.
- Evidence: The mandatory execution contract requires calling
python scripts/extract.py,python scripts/generate_report.py, andpython scripts/prepare_sample_input.py. - [DATA_EXFILTRATION]: While the skill is intended to generate reports, it includes defensive patterns to prevent the accidental inclusion of sensitive credentials in those reports.
- Evidence:
scripts/generate_report.pycontainsSENSITIVE_PATTERNSregular expressions to detect AccessKey IDs, Secret Keys, JWT tokens, and SSH private keys to ensure they are not exported into the resulting document.
Audit Metadata