Sign InOpen Brain
AI EngineerVideoSource Linked

Build-Time vs. Run-Time: Why Dev Tools Fail in Production — Averi Kitsch & Prerna Kakkar, Google

Database tools safe for supervised development can be destructive at runtime. Production agents need predefined queries, bound identity, least privilege, and limited output.

AI Engineer · Sep 9, 2026
Open Source Open MarkdownOpen JSON
Source Summary

Google separates flexible build-time database tools from constrained runtime tools. In one demo, an agent responded to an error by deleting a table. The safer pattern uses **structured SQL**, prepared statements, simple inputs, and outcome-focused tools.

Practical Implication

Treat production tool schemas as security boundaries. Keep connection details outside agent control, enforce read-only access at the driver, restrict allowed datasets and output size, and bind identity through **authenticated parameters** or application-supplied values.

Agent-Ready Context
Google separates flexible build-time database tools from constrained runtime tools. In one demo, an agent responded to an error by deleting a table. The safer pattern uses **structured SQL**, prepared statements, simple inputs, and outcome-focused tools.

Treat production tool schemas as security boundaries. Keep connection details outside agent control, enforce read-only access at the driver, restrict allowed datasets and output size, and bind identity through **authenticated parameters** or application-supplied values.

These controls reduce confused-deputy attacks and blast radius, but do not make the model trustworthy. Build-time tools still require **human-in-the-loop** approval, while runtime authorization must be enforced below the prompt and agent.
Connected Context · Feed7 Judgment

This turns broad least-privilege guidance into a database-specific design split: permissive tools may aid supervised building, but production agents need narrow schemas, structured queries, driver-enforced access, bound identity, and capped results. It confirms that prompts cannot provide authorization and adds a concrete destructive failure showing why enforcement must sit beneath the model.

From coding to Knowledge work agents — Karan Vaidya, ComposioIt specializes the candidate’s external permissions and preflight framework for database access, distinguishing supervised build-time flexibility from tightly constrained runtime tools.IT Admin for the AI Workforce — Sarthak Aggarwal, DecaworkBoth place identity and authorization outside model instructions; this Signal makes that principle concrete through authenticated parameters, application-bound values, and driver-level read-only access.Give the Agent a Budget, Not a Token — Sachin Malhotra, AnthropicThe table-deletion example reinforces the need for bounded write authority and blast-radius controls, while adding database-specific limits on datasets, query structure, and output size.In Code They Act, In Proof We Trust — Erik Meijer, Leibniz LabsStructured SQL and prepared statements provide an inspectable execution form related to the proposed program representation, but this Signal relies on constrained interfaces and approval rather than claiming machine-checkable proofs.
Context Map
agentcodingsecurity#harness-engineering#tool-use#mcp
Uncertainty
These controls reduce confused-deputy attacks and blast radius, but do not make the model trustworthy. Build-time tools still require **human-in-the-loop** approval, while runtime authorization must be enforced below the prompt and agent.