The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/generate_schematic.py uses subprocess.run to programmatically invoke scripts/generate_schematic_ai.py. While it uses a list-based argument structure to prevent shell injection, it passes user-provided prompts directly to the child process.
[EXTERNAL_DOWNLOADS]: The skill performs numerous network operations to fetch metadata and generate content. It queries authoritative sources including CrossRef (api.crossref.org), PubMed E-utilities (eutils.ncbi.nlm.nih.gov), and arXiv (export.arxiv.org). It also makes requests to OpenRouter (openrouter.ai) for AI-powered diagram generation and quality reviews.
[INDIRECT_PROMPT_INJECTION]: The skill has a significant attack surface for indirect prompt injection.
Ingestion points: Data enters the context via parallel-cli search/extract results (as described in SKILL.md Phase 2.5) and through LLM responses from OpenRouter in scripts/generate_schematic_ai.py.
Boundary markers: The instructions and scripts do not implement strict boundary markers or delimiters when interpolating external content into the agent's context or LLM prompts.
Capability inventory: The skill possesses capabilities for file writing (Write tool), network requests (requests library), and local command execution (subprocess.run).
Sanitization: There is no evidence of specific sanitization or filtering logic for the data retrieved from external web searches or LLM critiques.
[DYNAMIC_EXECUTION]: The skill generates and executes scripts locally to facilitate the diagram creation workflow. scripts/generate_schematic.py dynamically builds a command to execute another local Python script based on user input and project paths.

citation-management