The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands to verify skill packages, including python3, pytest, and npm. These commands are used to run a local audit script and execute the repository's test suite to ensure the new skill meets established standards.- [DYNAMIC_EXECUTION]: The skill generates new configuration files, markdown instructions, and scripts based on user requirements and then runs verification tests against these newly created artifacts. This is restricted to standard library usage and local repository paths.- [INDIRECT_PROMPT_INJECTION]: The skill functions as a code generator that ingests natural language descriptions to create new skill definitions, creating a potential surface for prompt injection in the generated output.
Ingestion points: The skill consumes user-provided triggers, phrasings, and task descriptions for the new skill package.
Boundary markers: The skill relies on a documented 'Skill Package Contract' to guide construction, but does not use explicit runtime delimiters to isolate user-provided data within the generated SKILL.md file.
Capability inventory: The skill has the capability to write to multiple repository directories (skills, commands, website) and execute shell commands via the agent.
Sanitization: The contract mandates validation of path arguments in generated scripts, but does not specify sanitization or escaping protocols for the natural language strings interpolated into frontmatter and body sections.

skill-forge