The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process and compile LaTeX documents, which serves as a significant attack surface for indirect prompt injection. LaTeX documents can contain malicious macros or commands that attempt to manipulate the AI agent's behavior or exfiltrate data.
Ingestion points: The skill accepts and processes LaTeX source code for academic papers, equations, and presentations (SKILL.md).
Boundary markers: The instructions do not specify the use of delimiters or provide the agent with guidance to ignore instructions embedded within the processed LaTeX source.
Capability inventory: The underlying latex dependency allows for file system interactions and, depending on configuration, command execution.
Sanitization: There is no evidence of input sanitization or instructions to use safety flags during compilation.
[DYNAMIC_EXECUTION]: The skill involves the runtime compilation of LaTeX documents. LaTeX is a powerful language that can perform complex operations, including reading local files (e.g., via \input) or executing shell commands (e.g., via \write18). If the compilation environment is not strictly sandboxed, this poses a risk of unauthorized system access.
[EXTERNAL_DOWNLOADS]: The skill documents a dependency on the latex system package and references the inbharatai/claude-skills repository on GitHub for its collection context.

latex-writer