code-then-execute-pattern.md

title	status	authors	based_on	category	source	tags
Code-Then-Execute Pattern	emerging	Nikola Balic (@nibzard)	DeepMind CaMeL (orig.) Luca Beurer-Kellner et al. (2025)	Tool Use & Environment	https://arxiv.org/abs/2506.08837	dsl sandbox program-synthesis auditability

Problem

Plan lists are opaque; we want full data-flow analysis and taint tracking.

Solution

Have the LLM output a sandboxed program or DSL script:

1. LLM writes code that calls tools and untrusted-data processors.
2. Static checker/Taint engine verifies flows (e.g., no tainted var to send_email.recipient).
3. Interpreter runs the code in a locked sandbox.

x = calendar.read(today)
y = QuarantineLLM.format(x)
email.write(to="john@acme.com", body=y)

How to use it

Complex multi-step agents like SQL copilots, software-engineering bots.

Trade-offs

• Pros: Formal verifiability; replay logs.
• Cons: Requires DSL design and static-analysis infra.

References

• Debenedetti et al., CaMeL (2025); Beurer-Kellner et al., §3.1 (5).