The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill reads external project files such as config.yaml and codebase artifacts to provide contextually relevant advice. This creates a surface for indirect prompt injection where instructions embedded in these files could attempt to influence the agent's behavior.
Ingestion points: Reads openspec/config.yaml and files specified in artifactPaths within the codebase.
Boundary markers: The skill does not explicitly instruct the agent to use delimiters or sanitization when processing these external files.
Capability inventory: The skill is authorized to use the openspec CLI via bash for reading and managing project specifications.
Sanitization: The skill relies on the agent to correctly handle file content as data rather than instructions, although it provides no explicit validation rules.
[COMMAND_EXECUTION]: The skill performs command execution using the openspec CLI tool through bash. It incorporates variables like store IDs and change names into these commands. The skill provides guidance on using double quotes to mitigate command injection risks, representing a standard use of specific project tooling.

openspec-explore