The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes project-specific data to provide context-aware analysis, which represents a potential injection surface for untrusted codebase content.
Ingestion points: Reads project context from openspec/config.yaml and accesses existing change artifacts.
Boundary markers: Includes instructions to treat external data as constraints rather than content to be reproduced.
Capability inventory: Uses the openspec CLI to manage artifacts and query repository status.
Sanitization: Relies on LLM reasoning to distinguish between exploration instructions and project data.
[DYNAMIC_EXECUTION]: The workflow relies on following instructions and templates provided dynamically by the openspec CLI tool.
The agent executes openspec instructions and applies the resulting templates to generate design documents.
Process logic allows for delegation to other specialized tools or skills as defined by the CLI's output metadata.

openspec-explore