The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external financial data provided by the user or via a JSON file.
Ingestion points: The scripts/altman_z.py script ingests data through command-line arguments (using argparse) or by reading a JSON file via the --input flag.
Boundary markers: The skill relies on structured input formats (floats and specific JSON keys) to delineate data from instructions.
Capability inventory: The script is limited to local arithmetic calculations and console output. It does not perform network operations, subprocess executions, or file system writes.
Sanitization: The script employs type validation via argparse (enforcing float types) and includes basic logic checks to ensure critical inputs like total_assets are positive. Due to the limited capabilities of the script, the risk of indirect injection is negligible.
[COMMAND_EXECUTION]: The skill uses a Python script to perform calculations. The script is static, uses standard libraries (argparse, json, sys), and does not take user-controlled input into shell commands or dangerous functions like eval() or exec().

algo-risk-altman-z