The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to perform codebase analysis and identify improvement opportunities within the user's local project. This creates an attack surface where malicious instructions embedded in code comments, documentation, or other project files could influence the agent's behavior during the tutorial.
Ingestion points: Local codebase files are analyzed in SKILL.md under the 'Tutorial phases' (Welcome and codebase analysis).
Boundary markers: No specific delimitation or 'ignore embedded instructions' warnings are provided for the analysis phase.
Capability inventory: The skill uses commands like /opsx:new and /opsx:apply to modify the local file system based on the analysis.
Sanitization: No sanitization or validation of the ingested codebase content is described.

openspec-onboard