The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands to create directories, execute a generated Python script for image creation, and perform self-checks using grep on the produced article files.
[DATA_EXPOSURE]: Reads internal repository data files including track_record.json (P&L history) and daily-picks.json (market scanner output) to populate the article content. This behavior is documented and central to the skill's purpose.
[DYNAMIC_EXECUTION]: Dynamically generates a Python script (generate_images.py) based on article data and then executes it using python3 to produce PNG infographics.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it processes external data from JSON files to influence its content generation.
Ingestion points: docs/api/daily-picks.json, landing/blog/blog_entries.json, and docs/backtesting/track_record.json.
Boundary markers: The skill employs a strict structural skeleton for the README.md and a list of "Voice gates" (formatting/content rules) to constrain output.
Capability inventory: File system write access, directory creation, and local execution of Python and shell commands.
Sanitization: No explicit sanitization or filtering of the ingested JSON data is performed before it is used in the text generation process.

draft-article