The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute the shinka_run CLI tool to perform batch mutations and evaluations of programs.
[INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of user feedback to update the evolutionary system prompt, which influences the code generation process.
Ingestion points: User feedback collected between batches is used to populate the evo.task_sys_msg parameter in SKILL.md.
Boundary markers: No explicit delimiters are used in the command-line examples; however, the skill suggests using --config-fname for long inputs to mitigate shell escaping issues.
Capability inventory: The shinka_run tool is capable of executing code, writing to the filesystem, and interacting with SQLite databases as described in SKILL.md.
Sanitization: There is no evidence of explicit sanitization or filtering of the user-provided feedback before it is passed to the CLI.
[DYNAMIC_EXECUTION]: The core functionality of the skill is to manage a framework that generates and executes program mutations.
Script generation + execution: The framework mutates initial.<ext> and executes it via evaluate.py to optimize performance.
Context: This behavior is the primary intended purpose of the skill for program evolution and discovery.

shinka-run