The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on the openspec CLI tool to manage 'change' directories and generate artifact instructions. All shell commands are scoped to the openspec binary as defined in the allowed-tools configuration.
[INDIRECT_PROMPT_INJECTION]: The skill ingests user descriptions and external dependency files to generate new planning artifacts. While this introduces an attack surface where malicious instructions could be embedded in the input data, the skill mitigates this by using structured templates and instructing the agent to clarify material ambiguities before proceeding.
[REMOTE_CODE_EXECUTION]: Although the skill executes shell commands, these are restricted to the local openspec tool. There are no patterns involving the download or execution of remote scripts via curl, wget, or similar utilities.

openspec-propose