The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill reads data from the local codebase and project configuration files to inform its thinking and exploration, which could contain instructions intended to influence the agent.
Ingestion points: Processes project background and conventions from openspec/config.yaml or config.yml, and reads arbitrary files while investigating the codebase.
Boundary markers: The agent is instructed to use these files as constraints and not to copy them, but there are no explicit delimiters or markers used to separate untrusted content from instructions.
Capability inventory: The skill is authorized to execute openspec CLI commands and read local files.
Sanitization: No evidence of sanitization, filtering, or validation of the content read from the filesystem.
[COMMAND_EXECUTION]: The skill executes openspec CLI commands via Bash to list store configurations, create new changes, and manage project artifacts. These executions are constrained to the openspec tool as specified in the metadata.

openspec-explore