The Agent Skills Directory

[SAFE]: The skill is designed for legitimate security research purposes, specifically reverse engineering. It provides instructional guidance and workflows for analyzing binary code and documenting findings within a Ghidra environment. No malicious patterns, data exfiltration, or unauthorized command execution were detected.
[INDIRECT_PROMPT_INJECTION]: The skill analyzes untrusted binary data (decompiled code, strings, and memory), which represents an inherent vulnerability surface for indirect prompt injection. Malicious binary content could theoretically contain instructions designed to mislead the agent's reasoning.
Ingestion points: Binary data is brought into the agent's context through tools like get-decompilation, get-strings, and read-memory (SKILL.md).
Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded commands within the binary data being analyzed.
Capability inventory: The skill allows the agent to modify the Ghidra database by renaming variables, changing data types, and adding comments or bookmarks (SKILL.md).
Sanitization: There is no explicit sanitization process mentioned for the data extracted from the binary targets.

deep-analysis