The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/blog_index.py script executes the hugo list all command using subprocess.run to synchronize draft statuses from the Hugo static site generator. While the command and its arguments are static, the execution relies on the hugo binary being present in the user's environment.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from local Markdown files (index.md) within a Hugo project directory. This content is processed and interpolated into the {{coverage}} variable within the references/idea-storm.md prompt template.
Ingestion points: The scripts/blog_index.py script recursively reads file content and frontmatter from the content/blog directory.
Boundary markers: The LLM prompt in references/idea-storm.md lacks explicit delimiters or instructions to ignore embedded commands around the interpolated {{coverage}} data.
Capability inventory: The skill has permissions to read local files and execute the hugo command via subprocess.
Sanitization: The script performs basic whitespace cleaning and metadata extraction but does not implement specific sanitization to strip potential LLM instructions from the processed blog content.
[DYNAMIC_EXECUTION]: The skill invokes system processes via subprocess.run. Although the command is hardcoded to hugo list all, the working directory is dynamically determined at runtime based on the detected project root or a user-supplied path, which is a common pattern for developer-oriented tools but involves dynamic context.

write-ideate