# Starter Kit for Testing Agentic Systems for Safety and Reliability

- **Record type**: Policy instrument
- **Jurisdiction**: Singapore
- **Instrument type**: Guidance
- **Status**: Guidance
- **Binding**: No — guidance, strategy or standard
- **Issuing body**: Infocomm Media Development Authority (IMDA)
- **Published**: 2026-10-14

## In brief

The Singapore starter kit for testing agentic AI is a non-binding guidance of Singapore, issued by the Infocomm Media Development Authority (IMDA). Its current status is guidance. Organisations testing agentic AI systems; presented as guidance, with no legal force of its own.

## Summary

A practical reference for organisations starting to test AI agents that reason, act on a user's behalf, use tools and access data. It sets out four steps (identify risks such as erroneous or unauthorised actions; design and run realistic tests without real-world harm; analyse results and operational logs; monitor and retest continuously), tests the agent's whole trajectory of reasoning and actions rather than only its output, and suggests metrics such as tool-use accuracy, handoff accuracy and human-intervention rates. It extends the agentic AI governance framework and builds on the LLM starter kit.

## Scope

Organisations testing agentic AI systems; presented as guidance, with no legal force of its own.

## Who it applies to

Developers, deployers and testers of AI agents.

## Key dates

Launched 14 October 2026 at the AI Assurance Exchange in Singapore.

## Notes on dates

Version, consultation window and voluntary status were not stated in the sources found; the earlier LLM starter kit describes itself as voluntary guidelines. Reported to draw on consultations with more than 20 deployers and testers (Google, AWS and Resaro named) and on the Global AI Assurance Sandbox, with case studies from Cohere, CyberSierra, Earlybird AI, KBTG and Knovel Engineering.

## Provenance

- **Permanent id**: AIPT-POL-0244 (https://aipolicytracker.org/id/AIPT-POL-0244; never changes or gets reused)
- **Record page**: https://aipolicytracker.org/policies/singapore-starter-kit-testing-agentic-systems
- **Official source**: https://www.smartcitiesworld.net/news/singapore-strengthens-ai-safety-and-resilience-13295
- **Source document**: Singapore strengthens AI safety and resilience
- **Published by**: Smart Cities World
- **Review status**: pending review
- **Confidence**: low
- **Facts last confirmed**: never confirmed against the official source
- **Retrieved**: 2026-10-21
- **Licence**: https://creativecommons.org/licenses/by/4.0/
- **Cite as**: AIPolicyTracker (2026). "Starter Kit for Testing Agentic Systems for Safety and Reliability" [AIPT-POL-0244, record version 1, not yet verified]. https://aipolicytracker.org/id/AIPT-POL-0244 (accessed 21 October 2026).

> This record is a structured summary with a link to the official text. It is not legal advice. Open the official source before relying on any date or duty. How current each record type must be is published at https://aipolicytracker.org/verification; what a record must carry at all is published at https://aipolicytracker.org/coverage.