fix: make permission callback e2e test more robust #485
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
- Use unique file paths (with UUID) to avoid conflicts with existing files - Explicitly instruct Claude to create a new file without reading first - Clean up test files before and after the test - Store both tool name and input data in callback invocations This fixes flaky test failures that occurred when /tmp/test.txt already existed, causing Claude to read the file first and potentially skip the write if content was identical. Claude-Generated-By: Claude Code (cli/claude-opus-4-5=100%) Claude-Steers: 0 Claude-Permission-Prompts: 8 Claude-Escapes: 0
The CLI auto-allows certain read-only commands (like 'echo') without consulting the SDK's can_use_tool callback. This caused the test to fail because the callback was never invoked. Changed the test to use 'touch' instead of 'echo', which is NOT in the CLI's read-only command list and requires permission approval. Also: - Use unique file paths (with UUID) to avoid conflicts - Clean up test files after test completes - Updated docstring to explain the read-only command behavior Claude-Generated-By: Claude Code (cli/claude-opus-4-5=100%) Claude-Steers: 9 Claude-Permission-Prompts: 27 Claude-Escapes: 0
antrewmorrison
approved these changes
Jan 16, 2026
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary
Fix the flaky
test_permission_callback_gets_callede2e test.Root Cause
The CLI auto-allows certain read-only commands (like
echo 'hello') without consulting the SDK'scan_use_toolcallback. The original test usedecho, which matched the CLI's read-only command pattern:When a command is classified as read-only, the CLI returns
behavior: 'allow'immediately and skips the SDK callback entirely.Fix
Changed the test to use
touchinstead ofecho. Thetouchcommand is NOT in the CLI's read-only command list, so it requires permission approval and correctly invokes the SDK callback.Changes
touchcommand instead ofecho(not auto-allowed)Testing
Ran the test 5+ times locally, all passed.