knowledge-eval

What problem does it solve?

This skill consolidates how Panopticon agents invoke Mission Control knowledge evaluation, eliminating ad-hoc client-side logic and ensuring consistent requests and governance.

Core Features & Use Cases

• Standardized requests: Centralizes POST /v1/knowledge/resolve calls across all agents.
• Smart defaults: Automatically fills agent_slug, API URLs, risk level, and result limits from environment variables.
• Consistent summaries: Returns a normalized evaluation summary with top evidence and constraints, alongside raw responses.
• Flexible filters: Supports optional tags, source_type, retrieval_mode, ranking_profile, and thresholds.
• Use cases: When orchestrating multiple agents, use this skill to ensure uniform knowledge evaluation and governance.

Quick Start

Run the knowledge-eval script from the skill directory with a sample task, agent slug, and risk level to generate an evaluation.