The Agent Skills Directory

[SAFE]: No malicious patterns or security vulnerabilities were detected in the skill instructions.
[COMMAND_EXECUTION]: The skill utilizes a restricted set of shell commands specifically for LaTeX document processing, including latexmk, pdflatex, and bibliography management tools. These commands are executed locally and are consistent with the skill's purpose.
[DATA_EXPOSURE]: The skill accesses specific local paths related to cloud storage (e.g., ~/Library/CloudStorage/) to discover research projects. This is a legitimate operational requirement for its intended use case and does not involve external exfiltration.
[DYNAMIC_EXECUTION]: Automatically generates a .latexmkrc configuration file if it is missing from a project directory. This behavior is used to standardize build settings and does not involve the execution of untrusted code from external sources.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data such as LaTeX source files and build logs to identify and fix errors. While this creates a theoretical attack surface, the risk is mitigated by the use of predefined fix patterns and the restriction of agent capabilities to a specific toolset.
Ingestion points: Reads .tex, .bib, and .log files from project directories.
Boundary markers: None explicitly defined for file content processing.
Capability inventory: Includes Write, Edit, and Bash execution for LaTeX-related tasks.
Sanitization: Fixes are applied based on rigid pattern matching (e.g., adding missing packages) rather than executing arbitrary instructions found in logs.

latex-health-check