skills/ascend-ai-coding/awesome-ascend-skills/external-cannbot-model-model-infer-parallel-impl/Gen Agent Trust Hub
external-cannbot-model-model-infer-parallel-impl
Pass
Audited by Gen Agent Trust Hub on Jul 7, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill ingests untrusted configuration data used to guide the model adaptation process.\n
- Ingestion points: The
parallel_configparameters in Step 1 ofSKILL.mdare sourced from user input or the main agent prompt.\n - Boundary markers: There are no explicit delimiters or instructions to ignore embedded commands within the configuration values.\n
- Capability inventory: The skill facilitates the execution of local scripts
infer.shandutils/weight_convert.sh.\n - Sanitization: The skill does not implement validation or sanitization of the configuration values before they are used in the implementation workflow.\n- [COMMAND_EXECUTION]: The implementation workflow includes instructions to execute local shell scripts
bash infer.shandbash utils/weight_convert.sh. These commands are standard for validating model changes and processing weights within the Ascend NPU development environment.
Audit Metadata