Execute an evaluation YAML produced by pmstack-eval against a real target system, running n_trials per task and reporting both pass@k and pass^k. Implements the running side of Anthropic's eval framework. HARD-STOPS with educative errors if no target is configured, no tasks are present, or required
Execute an evaluation YAML produced by pmstack-eval against a real target system, running n_trials per task and reporting both pass@k and pass^k. Implements the running side of Anthropic's eval framework. HARD-STOPS with educative errors if no target is configured, no tasks are present, or required env vars are missing — never simulates results. Use when the user asks to "run", "execute", or "scor
Esta entrada no publica ningún paquete de npm, así que Forge no tiene un árbol de dependencias para ella. Es una carencia de cobertura, no una afirmación de que no tenga dependencias.