The Agent Skills Directory

[SAFE] (SAFE): The skill contains no evidence of prompt injection, hardcoded credentials, malicious remote code execution, or unauthorized network activity. It functions correctly as a developer cheatsheet for legitimate tools.
[Indirect Prompt Injection] (SAFE): The skill documents tools that ingest external data, representing a theoretical attack surface.
Ingestion points: Commands such as strings, addr2line, and c++filt (via stdin) in both SKILL.md and references/cheatsheet.md process data from external files.
Boundary markers: No markers or warnings regarding untrusted content are included in the examples.
Capability inventory: The skill enables shell-based binary manipulation (e.g., objcopy, ar).
Sanitization: No input validation or sanitization steps are provided for the processed data.

binutils