The Agent Skills Directory

[COMMAND_EXECUTION]: Utilizes the openspec CLI for administrative tasks, status checks, and artifact generation. Command execution is constrained to the openspec namespace via frontmatter configuration.
[INDIRECT_PROMPT_INJECTION]: Processes data from user descriptions and CLI outputs to influence the generation of project artifacts.
Ingestion points: User-supplied change descriptions and the output of openspec instructions (SKILL.md).
Boundary markers: The skill specifies that context and rules provided by the CLI are constraints for the agent's logic and must not be included in final output files.
Capability inventory: File system write access for artifacts and execution of the openspec CLI.
Sanitization: Implements a naming convention (kebab-case) for new changes and utilizes CLI-provided templates for file structure.

openspec-propose