The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python script (scripts/soc_integrate.py) to handle Verilog parsing and code generation. This execution is performed using a runtime wrapper and is restricted to the tools provided within the skill package. Additionally, the entry point dynamically modifies the Python search path to resolve local dependencies.
[DATA_EXFILTRATION]: The skill performs file read and write operations to process RTL source code and save generated hardware descriptions, CSV reports, and JSON configuration files. These operations are local to the workspace and do not involve any network transmission.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it processes external Verilog files which could theoretically contain malicious instructions. However, the risk is minimal because the tool performs structural parsing and does not execute the contents of the files.
Ingestion points: Verilog source files are read in soc_integrate.py via the extract_modules and extract_map_from_top functions.
Boundary markers: None; the tool relies on the formal structural syntax of Verilog for data extraction.
Capability inventory: The skill possesses file read/write capabilities across its toolset to manage RTL assets.
Sanitization: The parser enforces strict ANSI-style port declaration rules and raises explicit errors for unsupported or complex syntax.

soc-integrate