The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes the openspec CLI via the Bash tool to perform project management tasks. This access is explicitly restricted to the openspec command namespace through the allowed-tools configuration, providing a layer of protection against arbitrary command injection.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from the local filesystem and CLI outputs—specifically instructions, templates, and artifact dependencies—to determine its next actions and the content of the files it writes. This represents the primary functional mechanism of the skill.
Ingestion points: Data is ingested from the JSON outputs of openspec instructions and openspec status commands, as well as the content of existing dependency files read from the disk.
Boundary markers: The instructions do not define specific delimiters for separating ingested content from the agent's internal logic, relying instead on the structured nature of the JSON responses.
Capability inventory: The skill has the capability to write files to local paths (resolvedOutputPath) and execute allowed openspec CLI sub-commands.
Sanitization: The skill follows the template and instruction fields provided by the CLI, using them as structured guidance for artifact generation.

openspec-ff-change