The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes vendor-specific openspec CLI commands and standard shell utilities (mkdir, mv) to manage the archiving process. These actions are restricted to the local project directory and align with the skill's documented functionality and intended use case.
[PROMPT_INJECTION]: The skill reads and evaluates local task lists and specification files, creating a surface for indirect prompt injection. The risk is assessed as low because the logic is focused on specific status parsing (e.g., checking for task markers) rather than open-ended instruction execution.
Ingestion points: The agent reads contents from tasks.md and various specification files in openspec/changes/ and openspec/specs/.
Boundary markers: Absent; the skill does not explicitly use delimiters to wrap content from external files.
Capability inventory: The skill can execute the openspec CLI and standard filesystem commands like mkdir and mv within the project scope.
Sanitization: No specific sanitization or escaping of external file content is implemented beyond the logic for parsing completion markers.

openspec-archive-change