The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of the openspec CLI, executing commands such as openspec new change, openspec status, and openspec instructions. These commands are used to initialize project structures, check status, and retrieve artifact generation requirements.
[EXTERNAL_DOWNLOADS]: The skill identifies a dependency on the openspec CLI, which must be installed on the host system to function. It does not attempt to download or install this dependency itself within the provided instructions.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by ingesting data from user prompts and external tool outputs to guide its behavior.
Ingestion points: Untrusted data enters via the AskUserQuestion tool (user-provided descriptions) and the JSON output of the openspec instructions command.
Boundary markers: There are no explicit boundary markers or XML-style delimiters used to isolate user input; however, CLI output is processed as structured JSON.
Capability inventory: The skill can execute shell commands via the openspec CLI and perform file write operations to paths specified by the tool output.
Sanitization: The skill relies on the agent to derive a safe 'kebab-case' name from user input, but it lacks explicit validation or escaping for the instructions retrieved from the external CLI.

openspec-propose