The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from project-level artifacts such as specification deltas, task lists, and design documents to perform verification logic. This content could be manipulated to include instructions that attempt to influence the agent's behavior during codebase analysis.
Ingestion points: Specification files (e.g., openspec/changes/<name>/specs/), tasks.md, and design.md accessed via contextFiles paths.
Boundary markers: Absent. The instructions do not specify the use of delimiters or 'ignore' commands when processing requirement text or scenario descriptions.
Capability inventory: The agent performs file system reads, executes shell commands via the openspec CLI, and conducts keyword-based searches across the codebase.
Sanitization: No explicit sanitization, filtering, or validation of the data extracted from external artifacts is performed before it is used to guide the agent's search and assessment tasks.
[COMMAND_EXECUTION]: The skill executes openspec CLI commands (list, status, instructions apply) to retrieve metadata and file paths. This is the intended functionality for interacting with the vendor's specialized tooling.

openspec-verify-change