The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/generate_schematic.py uses subprocess.run to execute a companion AI generation script (scripts/generate_schematic_ai.py). This implementation is safe as it avoids shell=True and passes arguments as a list, preventing shell injection vulnerabilities.
[EXTERNAL_DOWNLOADS]: The skill facilitates network requests to reputable academic and technology services, including CrossRef, NCBI (PubMed), arXiv, and OpenRouter. These operations are essential for the skill's stated purpose of fetching metadata and generating visual schematics.
[PROMPT_INJECTION]: The skill includes a surface for indirect prompt injection due to its reliance on external data sources. Maliciously crafted metadata in academic databases could theoretically influence the agent's context when ingested.
Ingestion points: scripts/extract_metadata.py, scripts/search_google_scholar.py, scripts/search_pubmed.py.
Boundary markers: None identified; external metadata is placed directly into the agent's workspace.
Capability inventory: The skill has access to Read, Write, Edit, and Bash tools, and uses subprocess.run and the requests library.
Sanitization: Standard field formatting is performed (e.g., BibTeX protection), but no explicit sanitization for LLM instruction injection is present.

citation-management