Skip to content

feat(security): Implement workspace isolation and task definition hiding during agent execution #132

Description

@pradeepvrd

Summary

When running DevOps Bench evaluations, evaluated agents with file system tools (or search capabilities) can scan workspace paths, discover task definitions (tasks/**/task.yaml), and read golden outputs directly.

Scope

  1. Execute agent subprocesses inside isolated ephemeral directories without mounting the tasks/ directory.
  2. Prevent direct file discovery of golden manifests and evaluation keys.

Ref: Internal bug b/552484058 (b/533473646)

Metadata

Metadata

Labels

No labels
No labels

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions