The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local script fdk/tools/medic.py using python3. It also instructs the agent to display and recommend shell commands generated by this script to the user for system repairs. While these are local operations within the framework, they represent the execution of logic based on the script's output.
[INDIRECT_PROMPT_INJECTION]: The skill processes project documentation, rules, and code to verify their state. This exposes the agent to untrusted data within those files which could be used to manipulate the health report or the suggested repair commands. 1. Ingestion points: Files including rules (harness-doctor), documentation, and code within the project directory are processed by the tool. 2. Boundary markers: The skill does not define specific delimiters or instructions to ignore embedded prompts within the files it evaluates. 3. Capability inventory: The skill has the capability to execute a local Python script and generate shell commands for user execution. 4. Sanitization: There is no explicit requirement to sanitize or filter the content read from project files before the results are presented to the agent or the user.

medic