Lessons Learned from the Consortium: Tool Use in Agent Systems
A NIST-led taxonomy for AI agent tools provides a structured vocabulary that Australian agencies evaluating or procuring agentic AI systems could apply directly to risk and capability assessment.
Key points
- NIST and CAISI have developed two draft taxonomies for AI agent tool use, covering functionality and constrained access patterns.
- The taxonomies aim to create shared vocabulary across the AI supply chain - useful for procurement, risk assessment, and incident reporting.
- Australia has no equivalent published taxonomy for AI agent tools; NIST's work may inform future Australian guidance or procurement frameworks.
Implications for Australian agencies
- Consider Agencies evaluating or procuring agentic AI systems may want to consider whether NIST's functionality and access-pattern taxonomies can structure their own capability and risk assessments.
- Monitor Policy and standards teams could monitor how NIST develops these taxonomies further, particularly given their relationship to NIST AI 600-1 and potential influence on international standards.
Implications are AI-generated. Starting points, not advice — see methodology for how they're framed.
View original source
Copied.
"Lessons Learned from the Consortium: Tool Use in Agent Systems"
Source: NIST – AI News (topic 2753736)
Published: 5 August 2025
URL: https://www.nist.gov/news-events/news/2025/08/lessons-learned-consortium-tool-use-agent-systems
NIST's CAISI, through an AISIC workshop with approximately 140 experts, has developed two draft taxonomies for tool use in AI agent systems. The first organises tools by function - perception, reasoning, and action - and the second categorises tools by access permission and environment trust level (read-only versus write, trusted versus untrusted). The taxonomies are intended to support transparency and communication across the AI supply chain, enabling developers, deployers, and users to describe agent capabilities, constrain risks, and report incidents using shared vocabulary. NIST invites public feedback and encourages adaptation to specific organisational needs.
Implications for Australian agencies:
- [Consider] Agencies evaluating or procuring agentic AI systems may want to consider whether NIST's functionality and access-pattern taxonomies can structure their own capability and risk assessments.
- [Monitor] Policy and standards teams could monitor how NIST develops these taxonomies further, particularly given their relationship to NIST AI 600-1 and potential influence on international standards.
Retrieved from SIMS, 18 July 2026.