The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes bash commands such as mkdir and mv to manage the filesystem. These commands are used to create archive directories and relocate change folders, which is the primary function of the skill. The operations are performed within the local project root.
[INDIRECT_PROMPT_INJECTION]: The skill reads and processes external data from files like tasks.md and spec.md, as well as JSON output from the openspec CLI, which creates a potential surface for indirect instructions to be processed by the agent.
Ingestion points: The skill ingests user-provided change names, the contents of local Markdown task files, and structured JSON data from tool outputs.
Boundary markers: There are no explicit instructions for the agent to use delimiters or ignore potential embedded commands when reading the contents of tasks.md or spec files.
Capability inventory: The skill possesses capabilities to execute shell commands (mkdir, mv) and trigger further automated workflows through the Task tool.
Sanitization: The instructions do not specify sanitization or validation logic for text parsed from the tasks.md file before it is used to generate warnings or summaries.

openspec-archive-change