The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes external scripts using subprocess.run within scripts/run_ai_alignment.py. It calls Python scripts (compile_latex_with_bibtex.py and convert_latex_to_word.py) located in the research-literature-review dependency skill directory to perform document rendering. This is a primary function of the skill and uses localized paths derived from the platform's skill installation environment.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface by processing content from untrusted LaTeX, BibTeX, and PDF files.
Ingestion points: Data is ingested from user-provided research documents in scripts/run_ai_alignment.py via parse_latex_document, parse_bib_file, and extract_pdf_text.
Boundary markers: The skill structures data into a JSON format (ai_alignment_input.json) for the agent context but lacks explicit instruction-isolation delimiters or 'ignore' warnings within the individual text fields.
Capability inventory: The skill has the capability to write files to the workspace and execute shell commands via subprocess.run.
Sanitization: The skill performs basic text truncation and regex-based sentence splitting but does not implement active filtering or sanitization of potential instructions embedded in the source research documents.

research-citation-check