The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use shell commands for generating scientific visualizations and compiling documents. * Evidence: It executes 'python scripts/generate_schematic.py' with AI-generated descriptions to produce image assets. * Evidence: It uses 'xelatex' and 'bibtex' to process LaTeX templates and bibliographies into final PDF reports.
[EXTERNAL_DOWNLOADS]: The skill relies on retrieving scientific literature and data from external web sources. * Evidence: Directs the agent to use 'WebFetch' for accessing PubMed and 'WebSearch' for general scientific research, although these tools are not listed in the frontmatter's allowed-tools.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to the processing of untrusted external literature. * Ingestion points: Scientific articles and web search content fetched via 'WebFetch' and 'WebSearch' (e.g., in SKILL.md sections 2 and 3). * Boundary markers: No explicit instructions or delimiters are provided to the agent to prevent the execution of instructions found within fetched scientific content. * Capability inventory: The agent has access to 'Bash' for command execution and 'Write' for file system operations. * Sanitization: The workflow lacks explicit sanitization or verification steps for data retrieved from external scientific domains.

hypothesis-generation