security: remove shell=True from ReAct agent bash tool#169
Merged
Yif-Yang merged 1 commit intoJul 26, 2026
Conversation
Parse the command with shlex.split and run with shell=False, and restrict the executable to a python/python3 allow-list. This removes shell metacharacter injection (the benchmark only needs Python to manipulate spreadsheets). Adds tests for the allow-list gate and non-interpretation of shell metacharacters.
6 tasks
There was a problem hiding this comment.
Pull request overview
This PR hardens the SpreadsheetBench ReAct agent’s bash tool by removing shell interpretation of agent-supplied command strings and restricting execution to Python, reducing shell-metacharacter injection risk while preserving required benchmark functionality.
Changes:
- Replaces
subprocess.run(..., shell=True)withshlex.split(...)+shell=Falsein_run_bash. - Adds an executable allow-list gate intended to restrict execution to
python/python3. - Adds tests covering disallowed commands, allowed Python execution, and non-interpretation of shell metacharacters.
Reviewed changes
Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.
| File | Description |
|---|---|
skillopt/envs/spreadsheetbench/react_agent.py |
Removes shell=True, tokenizes commands, and adds an executable allow-list before running subprocesses. |
tests/test_react_agent_no_shell.py |
Adds regression tests asserting the hardening behavior (allow-list + metacharacter non-interpretation). |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Comment on lines
+263
to
+266
| exe_name = os.path.basename(parts[0]).lower() | ||
| # Strip .exe suffix on Windows (python.exe → python) | ||
| exe_stem = exe_name.split(".")[0] | ||
| if exe_stem not in _CMD_ALLOW: |
This was referenced Jul 26, 2026
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
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.
What
Replace
shell=Truein the ReAct agent's bash tool withshlex.split+shell=False, and restrict the executable to a small allow-list (python,python3). The spreadsheet benchmark only needs Python to manipulate files, so this removes shell-metacharacter injection without losing functionality.File:
skillopt/envs/spreadsheetbench/react_agent.pySecurity guarantee (and limits)
Tests
tests/test_react_agent_no_shell.py:curl ...) is blockedpython -ccommand runs;) are passed as arguments, not interpreted, so a chained command does not executeContext
One of several small PRs replacing bulk PR #166, per reviewer feedback to split and scope. Full suite green except 6 pre-existing Windows-only path-separator failures that also fail on
main.