The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is authorized to execute openspec CLI commands via Bash (e.g., openspec list, openspec status, openspec new change) to discover project state, scaffold changes, and manage design artifacts. Execution is scoped to the openspec command namespace.
[INDIRECT_PROMPT_INJECTION]: The skill processes content from the local project environment which could contain malicious instructions designed to influence the agent's behavior.
Ingestion points: Reads project-specific background and rules from openspec/config.yaml, as well as existing design artifacts and the application codebase to map architecture.
Boundary markers: The skill contains instructions to treat file content as 'constraints to follow' rather than 'content to reproduce', which provides a conceptual boundary but lacks technical delimiters to prevent the agent from obeying embedded instructions.
Capability inventory: The skill possesses the ability to execute shell commands (restricted to the OpenSpec CLI) and write design artifacts (proposals, designs, specs) to the filesystem.
Sanitization: No explicit sanitization, validation, or escaping of the content read from the codebase or configuration files is described before the data is ingested by the agent.

openspec-explore