The Agent Skills Directory

[COMMAND_EXECUTION]: The skill interacts with the local environment using the openspec CLI via a restricted Bash tool scope (openspec:*). It uses these commands to manage the lifecycle of project proposals, check status, and retrieve artifact templates.
[DYNAMIC_EXECUTION]: The workflow allows for artifact creation to be delegated to specific sub-commands or skills if specified by the openspec instructions command. This is a core functional feature of the OpenSpec architecture for generating complex artifacts.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data in the form of JSON instructions and templates returned by the openspec CLI and reads artifact files from the filesystem.
Ingestion points: Data enters the context via openspec status --json and openspec instructions --json (SKILL.md).
Boundary markers: The skill does not explicitly define delimiters for CLI-provided instructions.
Capability inventory: The skill has capabilities to write files to disk and execute openspec CLI commands.
Sanitization: There is no explicit sanitization step for the instructions returned by the CLI, though the skill instructions emphasize that metadata like context and rules are for the agent's guidance and should not be copied into output files.

openspec-propose