The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes the openspec CLI tool via openspec list --json to retrieve the current status of project changes. This is an expected and documented integration with the compatible CLI mentioned in the metadata.
[PROMPT_INJECTION]: The skill is designed to ingest and process content from the user's codebase and OpenSpec artifacts (such as proposal.md or design.md). This establishes an indirect prompt injection surface where instructions hidden in those files could attempt to influence the agent. The risk is minimized by the skill's explicit constraints against writing application code and the requirement for user confirmation before creating any project artifacts. Ingestion points: Codebase files and OpenSpec change artifacts. Boundary markers: None present. Capability inventory: File system read access and execution of the openspec CLI; writing code or implementing features is explicitly forbidden. Sanitization: No specific sanitization or escaping of file content is defined.

openspec-explore