MalDet · Static Security Report
The repository JustVugg/colibri shows a medium vulnerability score of 51.0, which means there are several insecure coding patterns that an attacker could potentially exploit if the code is used in production. While some issues appear to be malicious attempts to hide credentials or evade security checks, many others look like ordinary mistakes made by developers who forgot to pin specific library versions or use weak passwords. The scan also identified 179 total findings, with several high-priority alerts pointing to hardcoded secrets and unsafe file access methods that need immediate attention before the code can be deployed safely.
Written by a language model running locally on the scanning machine, from the findings in this report. The risk levels above and below were decided by the scoring model, not by this summary.
MalDet scores two different questions separately, because they call for different responses: code can be badly written without being hostile, and hostile code is not always badly written.
| Detector | Looks for | High | Medium | Low | Total |
|---|---|---|---|---|---|
| bandit | Insecure coding patterns in Python source | 0 | 39 | 116 | 155 |
| yara | Signatures of malicious behaviour in source files | 10 | 0 | 7 | 17 |
| dep_checker | Vulnerable, unpinned or typosquatted dependencies | 0 | 5 | 2 | 7 |
15 of 179 findings were rated more likely genuine than noise by the classifier.
Identical findings are grouped, so a single rule firing many times takes one row. Ordered worst first. “Seen” is how many times that finding occurred.
| Severity | Detector | Finding | Seen | Example location |
|---|---|---|---|---|
| high | yara | YARA rule matched: detect_credential_harvester — References credential/secret material by name AND a deliberate outbound-send API in the same file - narrower than \'reads any env var\', which large legitimate codebases do constantly regardless of intent. Still not proof the two are connected (YARA c | 2 | /c/tools/convert_fp8_to_int4.py |
| high | yara | YARA rule matched: detect_anti_analysis — Checks for a debugger, VM, or sandbox before continuing - a common malware evasion technique, essentially never present in ordinary application code | 2 | /c/colibri.c |
| high | yara | YARA rule matched: detect_base64_exec — Encoded payload decoded and then executed | 2 | /c/openai_server.py |
| high | yara | YARA rule matched: detect_download_and_execute — Downloads a remote resource AND executes it - the actually dangerous combination, not just downloading alone | 4 | /c/tools/make_kimi_k3_ref.py |
| medium | dep_checker | Unpinned dependency: 'safetensors' has no version lock — supply chain risk | 1 | pyproject.toml |
| medium | dep_checker | Unpinned dependency: 'datasets' has no version lock — supply chain risk | 1 | pyproject.toml |
| medium | dep_checker | Unpinned dependency: 'tokenizers' has no version lock — supply chain risk | 1 | pyproject.toml |
| medium | dep_checker | Unpinned dependency: 'numpy' has no version lock — supply chain risk | 1 | pyproject.toml |
| medium | dep_checker | Unpinned dependency: 'huggingface_hub' has no version lock — supply chain risk | 1 | pyproject.toml |
| medium | bandit | Use of weak SHA1 hash for security. Consider usedforsecurity=False | 1 | /c/openai_server.py:2498 |
| medium | bandit | Audit url open for permitted schemes. Allowing use of file:/ or custom schemes is often unexpected. | 12 | /c/tools/repair_mtp_int8.py:53 |
| medium | bandit | Use of exec detected. | 1 | /colibri/_version.py:15 |
| medium | bandit | Unsafe Hugging Face Hub download without revision pinning in hf_hub_download() | 9 | /c/tools/convert_olmoe_merged.py:209 |
| medium | bandit | Unsafe Hugging Face Hub download without revision pinning in snapshot_download() | 4 | /c/tools/convert_olmoe.py:86 |
| medium | bandit | Unsafe Hugging Face Hub download without revision pinning in load_dataset() | 1 | /c/tools/fetch_benchmarks.py:51 |
| medium | bandit | Unsafe Hugging Face Hub download without revision pinning in from_pretrained() | 11 | /c/tools/make_qwen36_oracle.py:96 |
| low | bandit | Possible hardcoded password: '1' | 2 | /c/tools/make_glm53_tiny.py:85 |
| low | bandit | Possible hardcoded password: 'ok' | 1 | /c/doctor.py:723 |
| low | bandit | Possible hardcoded password: 'exact' | 1 | /c/tools/make_deepseek_v4_oracle.py:62 |
| low | bandit | Possible hardcoded password: '240' | 1 | /c/tools/mw_ur3_raw_adapter.py:62 |
| low | bandit | Possible hardcoded password: '0' | 5 | /c/tools/make_glm53_tiny.py:84 |
| low | bandit | Possible hardcoded password: 'None' | 8 | /c/tools/make_glm53_tiny.py:83 |
| low | bandit | Try, Except, Continue detected. | 1 | /c/tools/fetch_benchmarks.py:86 |
| low | bandit | Consider possible security implications associated with the subprocess module. | 19 | /c/tools/convert_glm53.py:236 |
| low | bandit | Starting a process with a partial executable path | 11 | /c/resource_plan.py:312 |
| low | bandit | subprocess call - check for execution of untrusted input. | 33 | /c/tools/blocksweep.py:37 |
| low | dep_checker | Loose version constraint: 'transformers>=4.40' — attacker could push malicious update | 1 | pyproject.toml |
| low | dep_checker | Loose version constraint: 'torch>=2.0' — attacker could push malicious update | 1 | pyproject.toml |
| low | yara | YARA rule matched: detect_downloader — Downloads a remote resource - common and low-signal alone; see detect_download_and_execute for the composite version | 7 | /c/download_fp8.py |
| low | bandit | Use of assert detected. The enclosed code will be removed when compiling to optimised byte code. | 34 | /c/tools/convert_fp8_to_int4.py:321 |