The Agent Skills Directory

[PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted data from external sources, such as binary files and firmware, which creates an indirect prompt injection surface. Malicious data within these files could attempt to influence the agent's instructions or bypass safety filters during the analysis process.
Ingestion points: Processes binaries, firmware, and disassembly output via Ghidra and IDA Pro.
Boundary markers: No specific boundary markers or 'ignore embedded instructions' warnings are mentioned for handling decompiled or disassembled output.
Capability inventory: The skill uses Bash, Write, and Edit tools, and describes the generation and execution of Python scripts (Ghidrathon/IDAPython).
Sanitization: No explicit sanitization or validation of the processed data is described before it is interpreted by the agent.
[COMMAND_EXECUTION]: The skill is configured to allow high-privilege tools including Bash. Its primary purpose involves generating and executing complex scripts for Ghidra and IDA Pro, which involves dynamic code execution at runtime.
[NO_CODE]: The skill documentation describes complex execution workflows but does not include any standalone executable scripts (e.g., .js, .py, .sh) within the provided files, relying instead on the agent to interact with pre-installed system dependencies like Ghidra and IDA Pro.

Ghidra/IDA Reverse Engineering Skill