alibabacloud-mining-attack-diagnosis
Fail
Audited by Gen Agent Trust Hub on Aug 21, 2026
Risk Level: CRITICALCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The script
scripts/_cli.pyusessubprocess.runto execute the officialaliyunCLI. This is a standard and necessary function for an Alibaba Cloud diagnostic skill. The command is constructed using a list of arguments, which is a secure implementation that prevents shell injection. - [SAFE]: The AV scan result
BitCoinMiner-AVforreferences/module2_alert_detail_ioc.mdis a false positive. The file contains reference data for miner family names (e.g., xmrig, kdevtmpfsi) and IOC patterns required for the skill to perform its intended detection and diagnosis functions. - [SAFE]: The skill is strictly read-only, as verified by code analysis of the invoked APIs (Describe, List, Query) and the restrictive RAM policies defined in
references/ram-policies.md. Mutating operations are explicitly prohibited. - [PROMPT_INJECTION]: The skill ingests untrusted alert data from Security Center APIs in
scripts/query_mining_alerts.py. Evidence chain: 1. Ingestion: SAS API response fields. 2. Boundaries: The skill outputs structured markdown with explicit headers. 3. Capability inventory: Read-only access to Alibaba Cloud APIs. 4. Sanitization: The skill filters data based on known mining keywords and masks sensitive identifiers in the final report.
Recommendations
- CRITICAL: 1 infected file(s) detected - DO NOT USE
Audit Metadata