The Compiler May Read It, the Agent May Not: Keeping Part of a Research Code Away from a Coding Agent

An agent reads external documents or tool responses that may contain malicious instructions.

What you would change

Replay attacks and legitimate lookalikes in an isolated copy of your agent workflow.

Replay attacks and legitimate lookalikes in an isolated copy of your agent workflow.

Compare against

Your existing checks with the same tasks, permissions, and tool responses.

1. Locate the paper's implementation and check its license, model support, data needs, and hardware requirements. Stop here if you cannot obtain the required artifacts.
2. Save a small set of representative product tasks and expected outcomes. Keep evaluation tasks out of any training or tuning, and choose quality and cost limits before running.
3. Run the baseline and the proposed change on identical inputs. Keep other settings fixed and repeat paired runs to expose variation.
4. Save per-task outputs, failures, timings, and all usage costs. Review regressions before expanding the test.

What to measure

Successful attacks missed, legitimate actions blocked, detection delay, and review cost.

When to keep it

Keep the change only if it meets your preselected quality and latency limits and improves the metric you care about after setup and operating costs. Otherwise retain the baseline. This is an exploration plan, not a validated recommendation.

recommendation

test

research area

safety and guardrails

technical change

Selective IP isolation from coding agents

published

2026-09-28

Teams running coding agents over proprietary code need selective-read isolation, but the paper demonstrates that standard sandboxing alone does not enforce it, so the approach requires workload-specific validation before production reliance.

1. claim 01Paperraft inference
   Read the brief interpretation

   The method replaces implicit trust in off-the-shelf harness controls (containers, permission rules, sandboxes) with an explicit classification of fifteen read routes, so that modules a physics solver needs at compile time are withheld from a coding agent. It costs engineering effort to audit and gate every read path in the build and agent toolchain, and adds architectural complexity, since the paper shows none of the three standard mechanisms can distinguish which program performs a read. Adoption can fail if any unclassified read route remains, if compiler-visible files leak through error messages, logs, or tool outputs, or if the harness is updated without re-auditing the route inventory.

   This is a Paperraft inference. It is not pinned to a verified PDF passage.

   Inspect the source