The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes local shell scripts and binaries for automation, specifically calling 'bash scripts/capture-screenshot.sh' and '~/.local/bin/bk-capture-screenshot'. The screenshot script employs Chrome headless with '--disable-web-security' and '--allow-file-access-from-files', which is a permissive configuration for rendering local visualizations.
[INDIRECT_PROMPT_INJECTION]: The skill reads and processes local project files ('index.md' and 'metadata.json'), which represents an attack surface for indirect prompt injection if the files contain malicious content meant to redirect agent actions. * Ingestion points: Content is read from 'docs/sims//index.md' and 'docs/sims//metadata.json' to extract titles and descriptions. * Boundary markers: Absent; the skill lacks delimiters or warnings to ignore embedded instructions in the ingested file content. * Capability inventory: The agent has file system write access (to 'mkdocs.yml', 'index.md', and 'TODO.txt') and shell command execution capabilities. * Sanitization: Absent; metadata and documentation text are reused directly in generated outputs without escaping or validation.
[DYNAMIC_EXECUTION]: The 'references/add-icons.md' component guides the agent in programmatically generating and injecting interactive JavaScript code into simulation files to add icons and click-based event logic using p5.js templates.

microsim-utils