The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use the Bash tool to execute npx openpersona commands. These commands are used to generate structural reports and evaluate persona configurations. This is a primary function of the skill and interfaces with the core OpenPersona framework provided by the vendor.
[INDIRECT_PROMPT_INJECTION]: The skill processes external persona files (persona.json, soul/.md) and agent-to-agent (A2A) handshake messages during semantic audits. This functionality represents an ingestion surface for untrusted data. 1. Ingestion points: local persona files and remote agent messages described in references/BLACK-BOX.md. 2. Boundary markers: The skill instructions specify reading structured packContent JSON and provide explicit markdown templates for output. 3. Capability inventory: Bash(npx openpersona:) and Read access to local persona files. 4. Sanitization: The skill implements strict evaluation rubrics and bias counters for the agent to follow during the assessment process. The risk is assessed as safe as these measures guide the agent toward analytical rather than executable interpretation of the content.

persona-evaluator