The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands to facilitate its workflow.
It executes a local binary at /opt/hatch/skills/skill-creator/bin/scaffold-connector-skill to generate skill templates.
It uses python3 -m py_compile to validate the syntax of generated Python scripts in the bin/ directory.
[INDIRECT_PROMPT_INJECTION]: The skill acts as a generator for new agent capabilities, creating an attack surface where user input influences the creation of instructions and scripts.
Ingestion points: User-provided descriptions and requirements for the new skill (SKILL.md).
Boundary markers: None explicitly defined to isolate user-supplied instructions during the generation process.
Capability inventory: Writing files to ~/workspace/skills/, executing a scaffolding script, and performing Python compilation (SKILL.md).
Sanitization: No explicit sanitization or content validation of the generated output is defined in the instructions.

skill_creator