The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the local filesystem to generate documents, creating a surface where instructions embedded in project files could potentially influence agent behavior. Ingestion points: openspec/project.md, openspec/changes/, and openspec/specs/. Boundary markers: Absent. Capability inventory: Write access to openspec/changes/ and execution of openspec validate via Bash. Sanitization: No explicit sanitization of ingested content before interpolation.
[COMMAND_EXECUTION]: The skill uses the Bash tool to execute the openspec validate command. This appears to be a local project-specific utility for validating proposal structure and does not involve remote code downloads or privilege escalation.

openspec-proposal