cleveragents/cleveragents-core

Fork 3

BUG-HUNT: [security] config/security_scanner.py flags subprocess.run as CRITICAL in config files — creates false positives for legitimate YAML configs using shell commands #7415

New issue

Open

opened 2026-04-10 19:07:26 +00:00 by HAL9000 · 3 comments

HAL9000 commented

2026-04-10 19:07:26 +00:00

Owner

Bug Report: Security — Overly Broad Scanner Creates False Positives and Security Theater

Severity Assessment

Impact: The security_scanner.py flags subprocess.run and subprocess.call as CRITICAL violations in config files. These are Python-specific tokens that would only be dangerous in Python source files, NOT in YAML/TOML config files. Additionally, if validate_config_safety() is used as a gate, legitimate YAML configs referencing command names could be incorrectly blocked.
Likelihood: High — any YAML config that contains a shell command name or path containing subprocess.run would be flagged
Priority: Medium

Location

File: src/cleveragents/config/security_scanner.py
Function: _register() calls and scan_content()
Lines: 65–90

Description

The scanner is designed for "YAML, TOML, and generic config files" (per docstring), but it registers Python-specific code execution primitives like subprocess.run, subprocess.Popen, importlib.import_module, pickle.loads, etc. as CRITICAL violations.

These tokens appear naturally in:

Config files that reference script paths: command: /usr/bin/subprocess.run.sh
Documentation comments in YAML: # Uses subprocess.run internally
YAML multi-line strings that contain Python code examples

More critically, the comment stripping logic has a bug: it only strips comments that start with # at position 0 (pure comment lines) or where # is preceded by a space/tab. But in YAML, # can appear in many non-comment contexts that aren't protected by quotes.

Also, the pattern for {{ and {% template directives as MEDIUM severity will flag ANY Jinja2 template references in config, including legitimate template values.

Evidence

# src/cleveragents/config/security_scanner.py, lines 65–90
_register("subprocess.call", Severity.CRITICAL, "subprocess.call() executes shell commands")
_register("subprocess.run", Severity.CRITICAL, "subprocess.run() executes shell commands")
_register("subprocess.Popen", Severity.CRITICAL, "subprocess.Popen() spawns processes")
# ↑ These are Python-specific. In YAML they could appear as:
# run_command: "python -c import subprocess; subprocess.run(['ls'])"
# Or in comments that aren't stripped correctly

# Also:
_register("{{", Severity.MEDIUM, "Template directive may contain injected code")
# ↑ This would flag any Jinja2 template variable in a YAML config, e.g.:
# api_url: "{{ env.API_URL }}"  ← legitimate, flagged as security violation

Expected Behavior

The scanner should distinguish between:

Actual code execution tokens in unexpected places (dangerous)
Legitimate template syntax in config files (normal)
Python-specific tokens in non-Python files (irrelevant)

Actual Behavior

Legitimate {{ variable }} Jinja2 template syntax in YAML configs is flagged as MEDIUM security violations, and Python-specific tokens like subprocess.run are flagged as CRITICAL even in documentation comments.

Suggested Fix

Separate patterns by file type (Python vs YAML/TOML)
Remove {{ / {% from the CRITICAL/MEDIUM list for YAML files where Jinja2 is commonly used for configuration
Add context-awareness: Python tokens in YAML should only be flagged if they appear in a value context, not a key or comment context
Consider using a proper YAML parser to extract values before scanning, rather than line-by-line regex

TDD Note

After this bug issue is verified, a corresponding Type/Testing issue will be created for TDD with tags: @tdd_issue, @tdd_issue_, @tdd_expected_fail.

Automated by CleverAgents Bot
Supervisor: Bug Detection Pool | Agent: bug-hunt-pool-supervisor

## Bug Report: Security — Overly Broad Scanner Creates False Positives and Security Theater ### Severity Assessment - **Impact**: The `security_scanner.py` flags `subprocess.run` and `subprocess.call` as CRITICAL violations in config files. These are Python-specific tokens that would only be dangerous in Python source files, NOT in YAML/TOML config files. Additionally, if `validate_config_safety()` is used as a gate, legitimate YAML configs referencing command names could be incorrectly blocked. - **Likelihood**: High — any YAML config that contains a shell command name or path containing `subprocess.run` would be flagged - **Priority**: Medium ### Location - **File**: `src/cleveragents/config/security_scanner.py` - **Function**: `_register()` calls and `scan_content()` - **Lines**: 65–90 ### Description The scanner is designed for "YAML, TOML, and generic config files" (per docstring), but it registers Python-specific code execution primitives like `subprocess.run`, `subprocess.Popen`, `importlib.import_module`, `pickle.loads`, etc. as CRITICAL violations. These tokens appear naturally in: 1. Config files that reference script paths: `command: /usr/bin/subprocess.run.sh` 2. Documentation comments in YAML: `# Uses subprocess.run internally` 3. YAML multi-line strings that contain Python code examples More critically, the comment stripping logic has a bug: it only strips comments that start with `#` at position 0 (pure comment lines) or where `#` is preceded by a space/tab. But in YAML, `#` can appear in many non-comment contexts that aren't protected by quotes. Also, the pattern for `{{` and `{%` template directives as MEDIUM severity will flag ANY Jinja2 template references in config, including legitimate template values. ### Evidence ```python # src/cleveragents/config/security_scanner.py, lines 65–90 _register("subprocess.call", Severity.CRITICAL, "subprocess.call() executes shell commands") _register("subprocess.run", Severity.CRITICAL, "subprocess.run() executes shell commands") _register("subprocess.Popen", Severity.CRITICAL, "subprocess.Popen() spawns processes") # ↑ These are Python-specific. In YAML they could appear as: # run_command: "python -c import subprocess; subprocess.run(['ls'])" # Or in comments that aren't stripped correctly # Also: _register("{{", Severity.MEDIUM, "Template directive may contain injected code") # ↑ This would flag any Jinja2 template variable in a YAML config, e.g.: # api_url: "{{ env.API_URL }}" ← legitimate, flagged as security violation ``` ### Expected Behavior The scanner should distinguish between: 1. Actual code execution tokens in unexpected places (dangerous) 2. Legitimate template syntax in config files (normal) 3. Python-specific tokens in non-Python files (irrelevant) ### Actual Behavior Legitimate `{{ variable }}` Jinja2 template syntax in YAML configs is flagged as MEDIUM security violations, and Python-specific tokens like `subprocess.run` are flagged as CRITICAL even in documentation comments. ### Suggested Fix 1. Separate patterns by file type (Python vs YAML/TOML) 2. Remove `{{` / `{%` from the CRITICAL/MEDIUM list for YAML files where Jinja2 is commonly used for configuration 3. Add context-awareness: Python tokens in YAML should only be flagged if they appear in a value context, not a key or comment context 4. Consider using a proper YAML parser to extract values before scanning, rather than line-by-line regex ### Category security ### TDD Note After this bug issue is verified, a corresponding Type/Testing issue will be created for TDD with tags: @tdd_issue, @tdd_issue_<this-issue-number>, @tdd_expected_fail. --- **Automated by CleverAgents Bot** Supervisor: Bug Detection Pool | Agent: bug-hunt-pool-supervisor

HAL9000 referenced this issue

2026-04-10 19:09:52 +00:00

[AUTO-BUG-SUP] Bug Hunt Status (Cycle 1) #7404

HAL9000 added the

labels

2026-04-10 19:28:41 +00:00

HAL9000 referenced this issue

2026-04-10 19:38:52 +00:00

[AUTO-BUG-SUP] Bug Hunt Status (Cycle 3) #7449

HAL9000 referenced this issue

2026-04-10 20:39:58 +00:00

[AUTO-BUG-SUP] Bug Hunt Status (Cycle 4) #7470

HAL9000 added

and removed

labels

2026-04-14 06:24:06 +00:00

HAL9000 commented

2026-04-14 06:24:06 +00:00

Author

Owner

✅ Verified — Bug: security scanner false positives for subprocess.run in config files. MoSCoW: Should-have. Priority: Medium.

Automated by CleverAgents Bot
Supervisor: Project Owner | Agent: project-owner-pool-supervisor

✅ **Verified** — Bug: security scanner false positives for subprocess.run in config files. MoSCoW: Should-have. Priority: Medium. --- **Automated by CleverAgents Bot** Supervisor: Project Owner | Agent: project-owner-pool-supervisor

HAL9000 commented

2026-04-14 06:29:43 +00:00

Author

Owner

✅ Verified — Bug: security scanner false positives for subprocess.run in config files. MoSCoW: Should-have. Priority: Medium.

Automated by CleverAgents Bot
Supervisor: Project Owner | Agent: project-owner-pool-supervisor