Frame high-stakes task prompts with accountability and self-checks.
npx skills add https://github.com/axiomantic/spellbook --skill emotional-stakes
Please help me install this Agent Skill. Skill: emotional-stakes Source: https://github.com/axiomantic/spellbook/tree/main/skills/emotional-stakes Command: npx skills add https://github.com/axiomantic/spellbook --skill emotional-stakes
Frames high-stakes task prompts with accountability to improve accuracy and truthfulness.
Select a professional persona, frame the task as substantive work, and state the stakes at the start.
High-intent search queries and answers about installing and using this skill.
Frame high-stakes prompts by selecting a professional persona, stating the stakes at the start, and structuring the task as substantive work with built-in self-checks to enforce accountability and truthfulness.
Persona-based task framing assigns a professional role to the AI, escalating risk awareness and enforcing self-monitoring to ensure the model handles risk-sensitive instructions with higher accountability.
You need to add self-checks to subagent prompts when executing performance-critical tasks, ethics reviews, or high-stakes decisions where automated self-monitoring is required to ensure accountability and accuracy.
Yes, you can use stake escalation for ethics reviews and AI safety prompts by explicitly stating the stakes at the start, which triggers self-directed framing and ensures accountability for high-stakes decisions.
Yes, framing high-stakes prompts works for performance-critical tasks across domains by applying in-task stake escalation and self-monitoring to ensure accountability and accuracy regardless of the specific domain.
Subagent prompts fail to maintain accountability when they lack explicit stake declarations and persona-based framing, preventing the model from triggering the self-directed framing and self-checks required for high-stakes decisions.
yu13130122297Identify uncertainties in research framing, datasets, metrics, and baselines.
bestdanAudit completed analysis directories for data integrity and reproducibility.
DTMC-marketplaceAssess AI systems for bias, fairness, and regulatory compliance.
UniquecreteDiagnose prompt weaknesses and rewrite prompts using the 5-Element Persona Framework.
