The Agent Skills Directory

[SAFE]: The skill implements a secure administrative process for preparing financial release packets and exhibits no malicious patterns.
Command Execution: The skill invokes a local Python script to process payment data within a restricted environment.
Data Access: It performs localized data processing and explicitly enforces a maker-checker protocol, stating it cannot perform financial transactions autonomously. It does not interact with sensitive system files or credentials.
Dynamic Loading: The script uses a relative path to import functions from a neighboring skill, which is a standard pattern for co-located skill dependencies within the environment.
[PROMPT_INJECTION]: The skill processes external data, which is a surface for indirect prompt injection.
Ingestion points: The script prepare_release_packet.py reads data from files provided via the --queue and --sanctions arguments.
Boundary markers: The output is structured as JSON, which delineates data fields, though no explicit instructional bypass warnings are present.
Capability inventory: The script is restricted to local file reading and printing to stdout; it has no network access or system command execution capabilities.
Sanitization: Input is parsed using the standard json library, and the skill includes logic to halt processing for payments that do not pass screening.

release-packet