The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/build.py uses subprocess.run to execute the latexmk build orchestrator. This is the skill's primary function, enabling the agent to compile LaTeX documents into PDFs. The implementation follows security best practices by passing arguments as a list and avoiding the use of shell=True for process invocation.
[PROMPT_INJECTION]: The skill contains an indirect prompt injection surface as outlined in references/error-triage.md, which instructs the agent to read and triage LaTeX .log files. Maliciously crafted LaTeX source files could generate logs containing hidden instructions intended to manipulate the agent's behavior during the error diagnosis phase.
Ingestion points: LaTeX log files (.log) are read and analyzed by the agent according to instructions in references/error-triage.md.
Boundary markers: There are no explicit boundary markers or instructions telling the agent to treat the log content as untrusted data or to ignore embedded commands.
Capability inventory: The skill includes scripts for executing shell commands (scripts/build.py) and deleting intermediate LaTeX files (scripts/clean.py).
Sanitization: The skill does not perform any sanitization or validation of the log file content before analysis.

latex