← Legacy view v2 (rp.*)

gen-verse/open-agentrl

https://github.com/Gen-Verse/Open-AgentRL.git · lang: python · LOC: · source: user_submitted

Quality
45.9
Grade D+
Security
9.9
Findings
38
0 critical · 3 high
Status
completed
May 16, 2026 02:01
medium: 20 info: 13 high: 3 low: 2
Top rules by occurrence
RuleSeverityCount
SEC001 Hardcoded Password critical 4
SEC005 Command Injection Risk high 4
SEC011 Unsafe PyTorch Model Loading medium 4
SEC014 SSL Verification Disabled medium 4
SEC020 Secret Printed to Logs high 4
SEC015 Insecure Randomness for Security medium 4
SEC007 Unsafe Deserialization medium 4
SEC012 ZipSlip — Archive Path Traversal medium 3
SEC016 LLM Prompt Injection — User Input in AI Prompt high 2
SEC013 Path Traversal — User Input in File Path high 2
First 38 findings (severity-sorted)
high SEC013 Path Traversal — User Input in File Path
verl/utils/reward_score/livecodebench/lcb_runner/evaluation/testing_util.py:131 · conf 0.80 
[SEC013] Path Traversal — User Input in File Path: User-controlled input used in file path without sanitization. Allows reading arbitrary files.
high SEC013 Path Traversal — User Input in File Path
verl/utils/reward_score/prime_code/testing_util.py:586 · conf 0.80 
[SEC013] Path Traversal — User Input in File Path: User-controlled input used in file path without sanitization. Allows reading arbitrary files.
high SEC016 LLM Prompt Injection — User Input in AI Prompt
verl/utils/reward_score/livecodebench/lcb_runner/prompts/self_repair.py:222 · conf 0.90 
[SEC016] LLM Prompt Injection — User Input in AI Prompt: User-supplied text is interpolated directly into an AI/LLM prompt (e.g. OpenAI, Anthropic, or local model). This is the AI equivalent of SQL i…
medium SEC001 Hardcoded Password
OSWorld-main/run_autoglm.py:531 · conf 0.30 
[SEC001] Hardcoded Password: Hardcoded password found in source code.
medium SEC001 Hardcoded Password
OSWorld-main/run_autoglm_v.py:565 · conf 0.30 
[SEC001] Hardcoded Password: Hardcoded password found in source code.
medium SEC001 Hardcoded Password
OSWorld-main/run_multienv_autoglm.py:538 · conf 0.30 
[SEC001] Hardcoded Password: Hardcoded password found in source code.
medium SEC005 Command Injection Risk
alfworld_eval.py:27 · conf 0.50 
[SEC005] Command Injection Risk: Unsafe shell execution or eval of user input.
medium SEC005 Command Injection Risk
coding_eval.py:15 · conf 0.50 
[SEC005] Command Injection Risk: Unsafe shell execution or eval of user input.
medium SEC005 Command Injection Risk
coding_rl.py:15 · conf 0.50 
[SEC005] Command Injection Risk: Unsafe shell execution or eval of user input.
medium SEC007 Unsafe Deserialization
verl/protocol.py:363 · conf 1.00 
[SEC007] Unsafe Deserialization: Unsafe deserialization can execute arbitrary code.
medium SEC007 Unsafe Deserialization
verl/utils/reward_score/livecodebench/lcb_runner/benchmarks/code_generation.py:68 · conf 1.00 
[SEC007] Unsafe Deserialization: Unsafe deserialization can execute arbitrary code.
medium SEC007 Unsafe Deserialization
verl/workers/rollout/sglang_rollout/utils.py:67 · conf 1.00 
[SEC007] Unsafe Deserialization: Unsafe deserialization can execute arbitrary code.
medium SEC011 Unsafe PyTorch Model Loading
train/osworld_train.py:566 · conf 1.00 
[SEC011] Unsafe PyTorch Model Loading: torch.load() uses pickle internally and can execute arbitrary code from untrusted model files.
medium SEC011 Unsafe PyTorch Model Loading
train/osworld_vlm_merge_preproc_shards.py:63 · conf 1.00 
[SEC011] Unsafe PyTorch Model Loading: torch.load() uses pickle internally and can execute arbitrary code from untrusted model files.
medium SEC011 Unsafe PyTorch Model Loading
verl/protocol.py:347 · conf 1.00 
[SEC011] Unsafe PyTorch Model Loading: torch.load() uses pickle internally and can execute arbitrary code from untrusted model files.
medium SEC012 ZipSlip — Archive Path Traversal
OSWorld-main/desktop_env/providers/docker/manager.py:88 · conf 1.00 
[SEC012] ZipSlip — Archive Path Traversal: Archive extraction without path validation allows writing files outside the target directory.
medium SEC012 ZipSlip — Archive Path Traversal
OSWorld-main/desktop_env/providers/virtualbox/manager.py:110 · conf 1.00 
[SEC012] ZipSlip — Archive Path Traversal: Archive extraction without path validation allows writing files outside the target directory.
medium SEC012 ZipSlip — Archive Path Traversal
OSWorld-main/desktop_env/providers/vmware/manager.py:185 · conf 1.00 
[SEC012] ZipSlip — Archive Path Traversal: Archive extraction without path validation allows writing files outside the target directory.
medium SEC014 SSL Verification Disabled
OSWorld-main/mm_agents/opencua/opencua_agent.py:449 · conf 1.00 
[SEC014] SSL Verification Disabled: SSL certificate verification is disabled, allowing man-in-the-middle attacks.
medium SEC014 SSL Verification Disabled
OSWorld-main/mm_agents/qwen3vl_agent_local.py:637 · conf 1.00 
[SEC014] SSL Verification Disabled: SSL certificate verification is disabled, allowing man-in-the-middle attacks.
medium SEC014 SSL Verification Disabled
OSWorld-main/mm_agents/uitars15_v1_local.py:778 · conf 1.00 
[SEC014] SSL Verification Disabled: SSL certificate verification is disabled, allowing man-in-the-middle attacks.
medium SEC016 LLM Prompt Injection — User Input in AI Prompt
verl/utils/reward_score/livecodebench/lcb_runner/prompts/code_generation.py:39 · conf 0.50 
[SEC016] LLM Prompt Injection — User Input in AI Prompt: User-supplied text is interpolated directly into an AI/LLM prompt (e.g. OpenAI, Anthropic, or local model). This is the AI equivalent of SQL i…
medium SEC017 Unbounded Input to LLM/External API
verl/utils/reward_score/livecodebench/lcb_runner/prompts/code_generation.py:39 · conf 0.80 
[SEC017] Unbounded Input to LLM/External API: User input is passed to an LLM or external AI API (OpenAI, Anthropic, etc.) without any visible length or size validation. This creates two risks: (1) Co…
low SEC006 XSS Risk
OSWorld-main/monitor/static/index.js:143 · conf 0.40 
[SEC006] XSS Risk: Direct HTML injection without sanitization.
low SEC017 Unbounded Input to LLM/External API
verl/utils/reward_score/livecodebench/lcb_runner/prompts/self_repair.py:222 · conf 0.30 
[SEC017] Unbounded Input to LLM/External API: User input is passed to an LLM or external AI API (OpenAI, Anthropic, etc.) without any visible length or size validation. This creates two risks: (1) Co…
info SEC001 Hardcoded Password
· conf 0.20 
[SEC001] Hardcoded Password (and 11 more): Same pattern found in 11 additional files. Review if needed.
info SEC005 Command Injection Risk
· conf 0.20 
[SEC005] Command Injection Risk (and 10 more): Same pattern found in 10 additional files. Review if needed.
info SEC007 Unsafe Deserialization
· conf 0.20 
[SEC007] Unsafe Deserialization (and 2 more): Same pattern found in 2 additional files. Review if needed.
info SEC011 Unsafe PyTorch Model Loading
· conf 0.20 
[SEC011] Unsafe PyTorch Model Loading (and 5 more): Same pattern found in 5 additional files. Review if needed.
info SEC014 SSL Verification Disabled
· conf 0.20 
[SEC014] SSL Verification Disabled (and 2 more): Same pattern found in 2 additional files. Review if needed.
info SEC015 Insecure Randomness for Security
· conf 0.20 
[SEC015] Insecure Randomness for Security (and 2 more): Same pattern found in 2 additional files. Review if needed.
info SEC015 Insecure Randomness for Security
alfworld_master/my_environments.py:330 · conf 0.25 
[SEC015] Insecure Randomness for Security: Weak PRNG used in security-sensitive context. Output is predictable.
info SEC015 Insecure Randomness for Security
OSWorld-main/desktop_env/providers/vmware/manager.py:70 · conf 0.25 
[SEC015] Insecure Randomness for Security: Weak PRNG used in security-sensitive context. Output is predictable.
info SEC015 Insecure Randomness for Security
train/utils.py:111 · conf 0.25 
[SEC015] Insecure Randomness for Security: Weak PRNG used in security-sensitive context. Output is predictable.
info SEC020 Secret Printed to Logs
· conf 0.20 
[SEC020] Secret Printed to Logs (and 13 more): Same pattern found in 13 additional files. Review if needed.
info SEC020 Secret Printed to Logs
sample/alfworld_utils.py:341 · conf 0.15 
[SEC020] Secret Printed to Logs: Debug or diagnostic code appears to print a credential-bearing value. This is a frequent AI-assisted coding failure: the helper exposes the exact value needed for tro…
info SEC020 Secret Printed to Logs
train/alfworld_train.py:590 · conf 0.15 
[SEC020] Secret Printed to Logs: Debug or diagnostic code appears to print a credential-bearing value. This is a frequent AI-assisted coding failure: the helper exposes the exact value needed for tro…
info SEC020 Secret Printed to Logs
train/osworld_train.py:542 · conf 0.15 
[SEC020] Secret Printed to Logs: Debug or diagnostic code appears to print a credential-bearing value. This is a frequent AI-assisted coding failure: the helper exposes the exact value needed for tro…

Reading from rp.scan + rp.finding + rp.rule (unified schema, R78 series). Legacy data path unchanged. Compare with /scan/67cbee14-0162-4d5c-96f7-35bb573129bc/.