The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the Bash(openspec:*) tool to interact with the local filesystem and the OpenSpec CLI. All command executions are scoped to the openspec command set, which is used to retrieve project status, change information, and artifact instructions.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data, establishing a surface for indirect prompt injection.
Ingestion points: Untrusted data enters the agent context through the JSON outputs of openspec status and openspec instructions, as well as through the content of dependency files read from the disk (e.g., existing spec files).
Boundary markers: The skill does not define specific delimiters or boundary markers for the content read from dependency files.
Capability inventory: The skill possesses the capability to execute openspec CLI commands and write new artifact files to the project directory.
Sanitization: The instructions explicitly direct the agent to treat context and rules fields as internal constraints rather than content to be copied, which serves as a prompt-level control against accidental inclusion of instructions in output artifacts.

openspec-continue-change