Scan report
ace-step / ACE-Step-1.5
Could an attacker exploit mistakes in this code?
Real issues worth reviewing before you rely on this code.
- High
- 0
- Medium
- 27
- Low
- 101
Sources: Bandit, Semgrep, pip-audit
Does this code look like it is trying to cause harm?
Serious issues. Review the high-severity findings before using it.
- High
- 14
- Medium
- 18
- Low
- 53
Sources: YARA, ClamAV, GuardDog, dependency checker
- Likely real
- 32
- Uncertain
- 10
- Likely noise
- 171
The risk classifier rates every finding for how likely it is to be a real issue. Likely noise still counts, just much less.
What we found, in plain English
This repository shows signs of both risky and potentially malicious code, so it should be treated with caution before use. The most serious issue is that several critical secrets like API keys and passwords are hardcoded directly into the source files, which means anyone who sees the code can steal them immediately. While some dependencies lack version locks, which creates supply chain risks, these issues appear to be unintentional rather than intentional attempts to harm others. The overall security score of 31.0 reflects medium-level vulnerabilities that could be exploited if the secrets are leaked or misused.
Written by a local language model from the findings below. The risk scores come from the classifier, not from this summary.
Detections by engine
Score over time
Overall score each time this repository was scanned.
Findings
Highest severity first. Select a finding to see the code and what it means.
-
- Engine
- GuardDog · Supply-chain attack heuristics
- Counts towards
- Malicious pattern
- File
- /cli.py line 15
- Likely real
- Yes, 96% confidence
Detects access to sensitive files (credentials, configs, keys) [threat-filesystem-read]
_project_root = os.path.dirname(_current_file) _env_path = os.path.join(_project_root, '.env') _env_example_path = os.path.join(_project_root, '.env.example') -
- Engine
- YARA · Malware signature rules
- Counts towards
- Malicious pattern
- File
- /start_gradio_ui_macos.sh
- Likely real
- Probably noise, 1% confidence
YARA rule matched: detect_download_and_execute — Downloads a remote resource AND executes it - the actually dangerous combination, not just downloading alone
-
- Engine
- YARA · Malware signature rules
- Counts towards
- Malicious pattern
- File
- /start_api_server_macos.sh
- Likely real
- Probably noise, 1% confidence
YARA rule matched: detect_download_and_execute — Downloads a remote resource AND executes it - the actually dangerous combination, not just downloading alone
-
- Engine
- YARA · Malware signature rules
- Counts towards
- Malicious pattern
- File
- /install_uv.sh
- Likely real
- Probably noise, 1% confidence
YARA rule matched: detect_download_and_execute — Downloads a remote resource AND executes it - the actually dangerous combination, not just downloading alone
-
- Engine
- YARA · Malware signature rules
- Counts towards
- Malicious pattern
- File
- /start_api_server.sh
- Likely real
- Probably noise, 1% confidence
YARA rule matched: detect_download_and_execute — Downloads a remote resource AND executes it - the actually dangerous combination, not just downloading alone
-
- Engine
- YARA · Malware signature rules
- Counts towards
- Malicious pattern
- File
- /start_gradio_ui_manual.sh
- Likely real
- Probably noise, 1% confidence
YARA rule matched: detect_download_and_execute — Downloads a remote resource AND executes it - the actually dangerous combination, not just downloading alone
-
- Engine
- YARA · Malware signature rules
- Counts towards
- Malicious pattern
- File
- /start_gradio_ui_macos_manual.sh
- Likely real
- Probably noise, 1% confidence
YARA rule matched: detect_download_and_execute — Downloads a remote resource AND executes it - the actually dangerous combination, not just downloading alone
-
- Engine
- YARA · Malware signature rules
- Counts towards
- Malicious pattern
- File
- /README.md
- Likely real
- Probably noise, 1% confidence
YARA rule matched: detect_download_and_execute — Downloads a remote resource AND executes it - the actually dangerous combination, not just downloading alone
-
- Engine
- YARA · Malware signature rules
- Counts towards
- Malicious pattern
- File
- /.claude/skills/acestep-docs/getting-started/README.md
- Likely real
- Probably noise, 1% confidence
YARA rule matched: detect_download_and_execute — Downloads a remote resource AND executes it - the actually dangerous combination, not just downloading alone
-
- Engine
- YARA · Malware signature rules
- Counts towards
- Malicious pattern
- File
- /.claude/skills/acestep-docs/guides/ENVIRONMENT_SETUP.md
- Likely real
- Probably noise, 1% confidence
YARA rule matched: detect_download_and_execute — Downloads a remote resource AND executes it - the actually dangerous combination, not just downloading alone
-
- Engine
- YARA · Malware signature rules
- Counts towards
- Malicious pattern
- File
- /start_gradio_ui.sh
- Likely real
- Probably noise, 1% confidence
YARA rule matched: detect_download_and_execute — Downloads a remote resource AND executes it - the actually dangerous combination, not just downloading alone
-
- Engine
- YARA · Malware signature rules
- Counts towards
- Malicious pattern
- File
- /install_uv.sh
- Likely real
- Probably noise, 0% confidence
YARA rule matched: detect_persistence_mechanism — Modifies a startup/scheduling location - legitimate application code essentially never does this
-
- Engine
- YARA · Malware signature rules
- Counts towards
- Malicious pattern
- File
- /.claude/skills/acestep-docs/guides/UPDATE_AND_BACKUP.md
- Likely real
- Probably noise, 0% confidence
YARA rule matched: detect_persistence_mechanism — Modifies a startup/scheduling location - legitimate application code essentially never does this
-
- Engine
- YARA · Malware signature rules
- Counts towards
- Malicious pattern
- File
- /.claude/skills/acestep-docs/guides/ENVIRONMENT_SETUP.md
- Likely real
- Probably noise, 0% confidence
YARA rule matched: detect_persistence_mechanism — Modifies a startup/scheduling location - legitimate application code essentially never does this
-
- Engine
- GuardDog · Supply-chain attack heuristics
- Counts towards
- Malicious pattern
- File
- /profile_inference.py line 124
- Likely real
- Yes, 96% confidence
Detects access to sensitive files (credentials, configs, keys) [threat-filesystem-read]
} env_file = os.path.join(PROJECT_ROOT, ".env") if os.path.exists(env_file): with open(env_file, "r", encoding="utf-8") as f: -
- Engine
- GuardDog · Supply-chain attack heuristics
- Counts towards
- Malicious pattern
- File
- /acestep/acestep_v15_pipeline.py line 18
- Likely real
- Yes, 96% confidence
Detects access to sensitive files (credentials, configs, keys) [threat-filesystem-read]
_project_root = os.path.dirname(os.path.dirname(_current_file)) _env_path = os.path.join(_project_root, ".env") _env_example_path = os.path.join(_project_root, ".env.example") if os.path.exists(_env_path): -
- Engine
- GuardDog · Supply-chain attack heuristics
- Counts towards
- Malicious pattern
- File
- /acestep/api_server.py line 173
- Likely real
- Yes, 96% confidence
Detects access to sensitive files (credentials, configs, keys) [threat-filesystem-read]
project_root = _get_project_root() env_path = os.path.join(project_root, ".env") if os.path.exists(env_path): load_dotenv(env_path, override=False) -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 44
- Likely real
- Yes, 54% confidence
Unpinned dependency: 'modelscope' has no version lock — supply chain risk
modelscope -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Yes, 54% confidence
Unpinned dependency: 'setuptools' has no version lock — supply chain risk
setuptools<72 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Yes, 54% confidence
Unpinned dependency: 'modelscope' has no version lock — supply chain risk
modelscope -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Yes, 54% confidence
Unpinned dependency: 'lycoris-lora' has no version lock — supply chain risk
lycoris-lora -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Yes, 54% confidence
Unpinned dependency: 'diskcache' has no version lock — supply chain risk
diskcache -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 62
- Likely real
- Yes, 54% confidence
Unpinned dependency: 'xxhash' has no version lock — supply chain risk
xxhash -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 23
- Likely real
- Yes, 54% confidence
Unpinned dependency: 'diffusers' has no version lock — supply chain risk
diffusers -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 48
- Likely real
- Yes, 54% confidence
Unpinned dependency: 'lycoris-lora' has no version lock — supply chain risk
lycoris-lora -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Yes, 54% confidence
Unpinned dependency: 'torchao' has no version lock — supply chain risk
torchao; platform_machine == 'aarch64' -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Yes, 54% confidence
Unpinned dependency: 'torchvision' has no version lock — supply chain risk
torchvision; sys_platform == 'darwin' and platform_machine == 'arm64' -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Yes, 54% confidence
Unpinned dependency: 'nano-vllm' has no version lock — supply chain risk
nano-vllm; sys_platform != 'darwin' or platform_machine != 'arm64' -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 43
- Likely real
- Yes, 54% confidence
Unpinned dependency: 'toml' has no version lock — supply chain risk
toml -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Yes, 54% confidence
Unpinned dependency: 'toml' has no version lock — supply chain risk
toml -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 4
- Likely real
- Uncertain, 50% confidence
Unpinned dependency: '--extra-index-url https://download.pytorch.org/whl/cu130' has no version lock — supply chain risk
--extra-index-url https://download.pytorch.org/whl/cu130 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 3
- Likely real
- Uncertain, 50% confidence
Unpinned dependency: '--extra-index-url https://download.pytorch.org/whl/cu128' has no version lock — supply chain risk
--extra-index-url https://download.pytorch.org/whl/cu128 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/training_v2/trainer_helpers.py line 339
- Likely real
- Uncertain, 49% confidence
Use of unsafe PyTorch load
338 if state_path.exists(): 339 state = torch.load( 340 str(state_path), map_location=module.device, weights_only=False 341 ) 342 epoch = state.get("epoch", 0) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/text_tasks/external_lm_model_discovery.py line 51
- Likely real
- Uncertain, 49% confidence
Audit url open for permitted schemes. Allowing use of file:/ or custom schemes is often unexpected.
50 try: 51 with request.urlopen(req, timeout=timeout_sec) as resp: 52 payload = json.loads(resp.read().decode("utf-8")) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /scripts/lora_data_prepare/gemini_caption.py line 184
- Likely real
- Probably noise, 22% confidence
Call to requests without timeout
183 url = f"{files_endpoint}/{file_name}" 184 response = requests.delete(url, headers=self._get_headers()) 185 return response.status_code == 204 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /scripts/lora_data_prepare/gemini_caption.py line 171
- Likely real
- Probably noise, 22% confidence
Call to requests without timeout
170 url = f"{files_endpoint}/{file_name}" 171 response = requests.get(url, headers=self._get_headers()) 172 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /scripts/lora_data_prepare/gemini_caption.py line 149
- Likely real
- Probably noise, 22% confidence
Call to requests without timeout
148 149 response = requests.post(upload_url, headers=upload_headers, data=file_data) 150 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /scripts/lora_data_prepare/gemini_caption.py line 125
- Likely real
- Probably noise, 22% confidence
Call to requests without timeout
124 upload_endpoint = self._get_endpoint("upload") 125 response = requests.post( 126 upload_endpoint, headers=init_headers, json=init_data 127 ) 128 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/model_downloader.py line 182
- Likely real
- Probably noise, 19% confidence
Unsafe Hugging Face Hub download without revision pinning in snapshot_download()
181 182 snapshot_download( 183 repo_id=repo_id, 184 local_dir=str(local_dir), 185 local_dir_use_symlinks="auto", 186 token=token, 187 ) 188 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/api/model_download.py line 54
- Likely real
- Probably noise, 19% confidence
Unsafe Hugging Face Hub download without revision pinning in snapshot_download()
53 54 snapshot_download( 55 repo_id=repo_id, 56 local_dir=download_dir, 57 local_dir_use_symlinks=False, 58 ) 59 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/example.py line 8
- Likely real
- Probably noise, 17% confidence
Unsafe Hugging Face Hub download without revision pinning in from_pretrained()
7 path = os.path.expanduser("~/huggingface/Qwen3-0.6B/") 8 tokenizer = AutoTokenizer.from_pretrained(path) 9 llm = LLM(path, enforce_eager=True, tensor_parallel_size=1) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /scripts/profile_vram.py line 384
- Likely real
- Probably noise, 17% confidence
Unsafe Hugging Face Hub download without revision pinning in from_pretrained()
383 print(f" Loading model weights...") 384 model = AutoModelForCausalLM.from_pretrained( 385 lm_path, 386 torch_dtype=torch.bfloat16, 387 trust_remote_code=True, 388 ) 389 model = model.to("cuda") -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /scripts/profile_vram.py line 329
- Likely real
- Probably noise, 17% confidence
Unsafe Hugging Face Hub download without revision pinning in from_pretrained()
328 tokenizer = AutoTokenizer.from_pretrained(encoder_path) 329 model = AutoModel.from_pretrained(encoder_path) 330 model = model.to("cuda").to(torch.bfloat16) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /scripts/profile_vram.py line 328
- Likely real
- Probably noise, 17% confidence
Unsafe Hugging Face Hub download without revision pinning in from_pretrained()
327 print(" Loading text encoder weights...") 328 tokenizer = AutoTokenizer.from_pretrained(encoder_path) 329 model = AutoModel.from_pretrained(encoder_path) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /scripts/profile_vram.py line 118
- Likely real
- Probably noise, 17% confidence
Unsafe Hugging Face Hub download without revision pinning in from_pretrained()
117 print(" Loading DiT model weights...") 118 model = AutoModel.from_pretrained( 119 model_path, 120 trust_remote_code=True, 121 attn_implementation="sdpa", 122 dtype=torch.bfloat16, 123 ) 124 model = model.to("cuda").to(torch.bfloat16) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/training_v2/model_loader.py line 315
- Likely real
- Probably noise, 17% confidence
Unsafe Hugging Face Hub download without revision pinning in from_pretrained()
314 tokenizer = AutoTokenizer.from_pretrained(str(text_path)) 315 encoder = AutoModel.from_pretrained(str(text_path)) 316 encoder = encoder.to(device=device, dtype=dtype) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/training_v2/model_loader.py line 314
- Likely real
- Probably noise, 17% confidence
Unsafe Hugging Face Hub download without revision pinning in from_pretrained()
313 dtype = _resolve_dtype(precision) 314 tokenizer = AutoTokenizer.from_pretrained(str(text_path)) 315 encoder = AutoModel.from_pretrained(str(text_path)) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/training_v2/model_loader.py line 234
- Likely real
- Probably noise, 17% confidence
Unsafe Hugging Face Hub download without revision pinning in from_pretrained()
233 result["text_tokenizer"] = AutoTokenizer.from_pretrained(str(text_path)) 234 text_enc = AutoModel.from_pretrained(str(text_path)) 235 text_enc = text_enc.to(device=device, dtype=dtype) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/training_v2/model_loader.py line 233
- Likely real
- Probably noise, 17% confidence
Unsafe Hugging Face Hub download without revision pinning in from_pretrained()
232 if text_path.is_dir(): 233 result["text_tokenizer"] = AutoTokenizer.from_pretrained(str(text_path)) 234 text_enc = AutoModel.from_pretrained(str(text_path)) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/training_v2/model_loader.py line 159
- Likely real
- Probably noise, 17% confidence
Unsafe Hugging Face Hub download without revision pinning in from_pretrained()
158 try: 159 model = AutoModel.from_pretrained( 160 str(model_dir), 161 trust_remote_code=True, 162 attn_implementation=attn_impl, 163 dtype=dtype, 164 ) 165 print(f"[OK] Model loaded with attn_implementation={attn_impl}") -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/config.py line 24
- Likely real
- Probably noise, 17% confidence
Unsafe Hugging Face Hub download without revision pinning in from_pretrained()
23 assert 1 <= self.tensor_parallel_size <= 8 24 self.hf_config = AutoConfig.from_pretrained(self.model) 25 self.max_model_len = min(self.max_model_len, self.hf_config.max_position_embeddings) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/core/generation/handler/init_service_loader.py line 175
- Likely real
- Probably noise, 17% confidence
Unsafe Hugging Face Hub download without revision pinning in from_pretrained()
174 logger.info(f"[initialize_service] Attempting to load model with attention implementation: {candidate}") 175 self.model = AutoModel.from_pretrained( 176 model_checkpoint_path, 177 trust_remote_code=True, 178 attn_implementation=candidate, 179 dtype=self.dtype, 180 ) 181 attn_implementation = candidate -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/core/generation/handler/init_service_loader_components.py line 94
- Likely real
- Probably noise, 17% confidence
Unsafe Hugging Face Hub download without revision pinning in from_pretrained()
93 94 self.text_tokenizer = AutoTokenizer.from_pretrained(text_encoder_path) 95 self.text_encoder = AutoModel.from_pretrained(text_encoder_path) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/core/generation/handler/init_service_loader_components.py line 95
- Likely real
- Probably noise, 17% confidence
Unsafe Hugging Face Hub download without revision pinning in from_pretrained()
94 self.text_tokenizer = AutoTokenizer.from_pretrained(text_encoder_path) 95 self.text_encoder = AutoModel.from_pretrained(text_encoder_path) 96 if not self.offload_to_cpu: -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/llm_inference.py line 404
- Likely real
- Probably noise, 17% confidence
Unsafe Hugging Face Hub download without revision pinning in from_pretrained()
403 try: 404 self.llm = AutoModelForCausalLM.from_pretrained(model_path, trust_remote_code=True) 405 if not self.offload_to_cpu: -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/llm_inference.py line 612
- Likely real
- Probably noise, 17% confidence
Unsafe Hugging Face Hub download without revision pinning in from_pretrained()
611 # TODO: load tokenizer too slow, not found solution yet 612 llm_tokenizer = AutoTokenizer.from_pretrained(full_lm_model_path, use_fast=True) 613 logger.info(f"5Hz LM tokenizer loaded successfully in {time.time() - start_time:.2f} seconds") -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/llm_inference.py line 4193
- Likely real
- Probably noise, 17% confidence
Unsafe Hugging Face Hub download without revision pinning in from_pretrained()
4192 start_time = time.time() 4193 self._hf_model_for_scoring = AutoModelForCausalLM.from_pretrained( 4194 model_path, 4195 trust_remote_code=True, 4196 torch_dtype=self.dtype 4197 ) 4198 load_time = time.time() - start_time -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/llm_inference.py line 4231
- Likely real
- Probably noise, 17% confidence
Unsafe Hugging Face Hub download without revision pinning in from_pretrained()
4230 start_time = time.time() 4231 self._hf_model_for_scoring = AutoModelForCausalLM.from_pretrained( 4232 model_path, 4233 trust_remote_code=True, 4234 torch_dtype=self.dtype 4235 ) 4236 load_time = time.time() - start_time -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/engine/llm_engine.py line 45
- Likely real
- Probably noise, 17% confidence
Unsafe Hugging Face Hub download without revision pinning in from_pretrained()
44 else: 45 self.tokenizer = AutoTokenizer.from_pretrained(config.model, use_fast=True) 46 config.eos = self.tokenizer.eos_token_id -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/text_tasks/external_lm_model_discovery.py line 96
- Likely real
- Yes, 95% confidence
Possible hardcoded password: 'ollama'
95 96 if provider_token == "ollama": 97 if "/v1/" in root: -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/text_tasks/passphrase_store.py line 13
- Likely real
- Yes, 95% confidence
Possible hardcoded password: 'acestep.external_lm'
12 13 EXTERNAL_LM_SECRET_SERVICE = "acestep.external_lm" 14 EXTERNAL_LM_SECRET_USERNAME = "external_lm_store_passphrase" -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/text_tasks/external_lm_providers.py line 78
- Likely real
- Yes, 95% confidence
Possible hardcoded password: 'ACESTEP_OLLAMA_SECRET_PATH'
77 api_key_required=False, 78 secret_path_env="ACESTEP_OLLAMA_SECRET_PATH", 79 secret_file_name="ollama_api_key.enc", 80 base_url_presets=( 81 ("Local Ollama", "http://127.0.0.1:11434/v1/chat/completions"), 82 ("Localhost alias", "http://localhost:11434/v1/chat/completions"), 83 ), 84 ), 85 } 86 87 88 def get_external_provider_profile(provider: str | None) -> ExternalProviderProfile: 89 """Return the provider profile for a provider identifier. 90 91 Raises: 92 ValueError: If ``provider`` is not a supported external provider. 93 """ -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/text_tasks/external_lm_providers.py line 64
- Likely real
- Yes, 95% confidence
Possible hardcoded password: 'ACESTEP_ANTHROPIC_SECRET_PATH'
63 api_key_required=True, 64 secret_path_env="ACESTEP_ANTHROPIC_SECRET_PATH", 65 secret_file_name="anthropic_api_key.enc", 66 base_url_presets=( 67 ("Anthropic messages", "https://api.anthropic.com/v1/messages"), 68 ), 69 ), 70 "ollama": ExternalProviderProfile( 71 provider_id="ollama", 72 label="Ollama", 73 protocol="openai_chat", 74 default_model="qwen3:4b", 75 default_base_url="http://127.0.0.1:11434/v1/chat/completions", 76 api_key_env="ACESTEP_OLLAMA_API_KEY", 77 api_key_required=False, 78 secret_path_env="ACESTEP_OLLAMA_SECRET_PATH", -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/text_tasks/external_lm_providers.py line 50
- Likely real
- Yes, 95% confidence
Possible hardcoded password: 'ACESTEP_OPENAI_SECRET_PATH'
49 api_key_required=True, 50 secret_path_env="ACESTEP_OPENAI_SECRET_PATH", 51 secret_file_name="openai_api_key.enc", 52 base_url_presets=( 53 ("OpenAI chat completions", "https://api.openai.com/v1/chat/completions"), 54 ), 55 ), 56 "claude": ExternalProviderProfile( 57 provider_id="claude", 58 label="Anthropic Claude", 59 protocol="anthropic_messages", 60 default_model="claude-3-7-sonnet-latest", 61 default_base_url="https://api.anthropic.com/v1/messages", 62 api_key_env="ACESTEP_ANTHROPIC_API_KEY", 63 api_key_required=True, 64 secret_path_env="ACESTEP_ANTHROPIC_SECRET_PATH", -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/text_tasks/external_lm_providers.py line 35
- Likely real
- Yes, 95% confidence
Possible hardcoded password: 'ACESTEP_GLM_SECRET_PATH'
34 api_key_required=True, 35 secret_path_env="ACESTEP_GLM_SECRET_PATH", 36 secret_file_name="glm_api_key.enc", 37 base_url_presets=( 38 ("Standard chat endpoint", "https://api.z.ai/api/paas/v4/chat/completions"), 39 ("Coding endpoint", "https://api.z.ai/api/coding/paas/v4/chat/completions"), 40 ), 41 ), 42 "openai": ExternalProviderProfile( 43 provider_id="openai", 44 label="OpenAI", 45 protocol="openai_chat", 46 default_model="gpt-4o-mini", 47 default_base_url="https://api.openai.com/v1/chat/completions", 48 api_key_env="ACESTEP_OPENAI_API_KEY", 49 api_key_required=True, 50 secret_path_env="ACESTEP_OPENAI_SECRET_PATH", -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/text_tasks/passphrase_store.py line 14
- Likely real
- Yes, 95% confidence
Possible hardcoded password: 'external_lm_store_passphrase'
13 EXTERNAL_LM_SECRET_SERVICE = "acestep.external_lm" 14 EXTERNAL_LM_SECRET_USERNAME = "external_lm_store_passphrase" 15 # Backward-compatible GLM aliases are kept for callers that still import them directly. -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/ui/gradio/events/results/session_artifacts.py line 22
- Likely real
- Yes, 95% confidence
Possible hardcoded password: 'lyric_token_ids'
21 "context_latents": "context_latents", 22 "lyric_token_idss": "lyric_token_ids", 23 } 24 REQUIRED_ARTIFACT_KEYS = tuple(ARTIFACT_FIELD_MAP.values()) 25 26 27 def persist_sample_session_artifacts( 28 extra_outputs: dict[str, Any], 29 sample_idx: int, -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/text_tasks/external_lm_model_discovery.py line 92
- Likely real
- Yes, 95% confidence
Possible hardcoded password: 'anthropic_messages'
91 provider_token = (provider or "").strip().lower() 92 if protocol_token == "anthropic_messages" and "/models" not in root and "/v1/" in root: 93 prefix = root.split("/v1/", 1)[0] -
- Engine
- GuardDog · Supply-chain attack heuristics
- Counts towards
- Malicious pattern
- File
- /acestep/api_server.py line 193
- Likely real
- Yes, 93% confidence
Detects reading of environment variables (credential access, often contains secrets) [threat-runtime-environment-read]
# API Key authentication (from environment variable) api_key = os.getenv("ACESTEP_API_KEY", None) set_api_key(api_key) -
- Engine
- GuardDog · Supply-chain attack heuristics
- Counts towards
- Malicious pattern
- File
- /acestep/handler.py line 9
- Likely real
- Yes, 93% confidence
Detects reading of environment variables (credential access, often contains secrets) [threat-runtime-environment-read]
# Disable tokenizers parallelism to avoid fork warning os.environ["TOKENIZERS_PARALLELISM"] = "false" import threading -
- Engine
- GuardDog · Supply-chain attack heuristics
- Counts towards
- Malicious pattern
- File
- /acestep/api/server_cli.py line 38
- Likely real
- Yes, 93% confidence
Detects reading of environment variables (credential access, often contains secrets) [threat-runtime-environment-read]
type=str, default=os.getenv("ACESTEP_API_KEY", None), help="API key for authentication (default from ACESTEP_API_KEY)", ) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/text_tasks/passphrase_store.py line 18
- Likely real
- Yes, 91% confidence
Possible hardcoded password: 'secret-tool'
17 GLM_SECRET_USERNAME = EXTERNAL_LM_SECRET_USERNAME 18 _SECRET_TOOL_PATH = "secret-tool" 19 _SECRET_TOOL_TIMEOUT_SEC = 5 -
- Engine
- GuardDog · Supply-chain attack heuristics
- Counts towards
- Malicious pattern
- File
- /acestep/api/runtime_helpers.py line 71
- Likely real
- Yes, 83% confidence
Detects fully silent process execution (suppressing all output channels) [threat-process-spawn-silent]
], stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL, ) -
- Engine
- GuardDog · Supply-chain attack heuristics
- Counts towards
- Malicious pattern
- File
- /acestep/api/runtime_helpers.py line 72
- Likely real
- Yes, 83% confidence
Detects fully silent process execution (suppressing all output channels) [threat-process-spawn-silent]
stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL, ) return f"http://localhost:{tensorboard_port}" -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/core/scoring/dit_alignment.py line 230
- Likely real
- Uncertain, 45% confidence
Possible hardcoded password: ''
229 # Incomplete UTF-8 sequence, this token doesn't complete a character 230 token_text = "" 231 else: -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/core/scoring/dit_alignment.py line 233
- Likely real
- Uncertain, 45% confidence
Possible hardcoded password: ''
232 # Edge case: current decode is shorter (shouldn't happen normally) 233 token_text = "" 234 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/text_tasks/passphrase_store.py line 53
- Likely real
- Uncertain, 45% confidence
Possible hardcoded password: ''
52 53 if passphrase == "": 54 return False, "Passphrase cannot be empty." -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/text_tasks/secure_secret_store.py line 68
- Likely real
- Uncertain, 45% confidence
Possible hardcoded password: ''
67 """Encrypt and store a secret value.""" 68 if secret == "": 69 raise SecretStoreError("Secret cannot be empty.") -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/text_tasks/secure_secret_store.py line 73
- Likely real
- Uncertain, 45% confidence
Possible hardcoded password: ''
72 return 73 if passphrase == "": 74 raise SecretStoreError("Passphrase cannot be empty.") -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/text_tasks/secure_secret_store.py line 101
- Likely real
- Uncertain, 45% confidence
Possible hardcoded password: ''
100 raise SecretStoreError(f"Secret not found at: {self.secret_path}") 101 if passphrase == "": 102 raise SecretStoreError("Passphrase cannot be empty.") -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/ui/gradio/events/training/lokr_training.py line 239
- Likely real
- Probably noise, 28% confidence
Try, Except, Continue detected.
238 epoch_num = int(d.split("_")[1]) 239 except Exception: 240 continue 241 checkpoints.append((epoch_num, d)) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/ui/gradio/events/training/lokr_training.py line 291
- Likely real
- Probably noise, 28% confidence
Try, Except, Continue detected.
290 int(d.split("_")[1]) 291 except Exception: 292 continue 293 checkpoint_names.append(d) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/constrained_logits_processor.py line 1125
- Likely real
- Probably noise, 28% confidence
Try, Except, Continue detected.
1124 1125 except Exception: 1126 continue 1127 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/constrained_logits_processor.py line 573
- Likely real
- Probably noise, 28% confidence
Try, Except, Continue detected.
572 logger.debug(f"Skipping audio code token {token_id} with invalid code value {code_value} (max: {MAX_AUDIO_CODE})") 573 except Exception: 574 continue 575 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/gpu_config.py line 611
- Likely real
- Probably noise, 14% confidence
Consider possible security implications associated with the subprocess module.
610 try: 611 import subprocess 612 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/text_tasks/passphrase_store.py line 7
- Likely real
- Probably noise, 14% confidence
Consider possible security implications associated with the subprocess module.
6 import shutil 7 import subprocess 8 from pathlib import Path -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/text_tasks/secure_secret_store.py line 9
- Likely real
- Probably noise, 13% confidence
Consider possible security implications associated with the subprocess module.
8 import shutil 9 import subprocess 10 import sys -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/api/runtime_helpers.py line 7
- Likely real
- Probably noise, 13% confidence
Consider possible security implications associated with the subprocess module.
6 import os 7 import subprocess 8 import sys -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/audio_utils.py line 14
- Likely real
- Probably noise, 13% confidence
Consider possible security implications associated with the subprocess module.
13 import os 14 import subprocess 15 import hashlib -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /scripts/check_gpu.py line 17
- Likely real
- Probably noise, 13% confidence
Consider possible security implications associated with the subprocess module.
16 import sys 17 import subprocess 18 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/gpu_config.py line 613
- Likely real
- Probably noise, 12% confidence
Starting a process with a partial executable path
612 613 result = subprocess.run( 614 ["sysctl", "-n", "hw.memsize"], 615 capture_output=True, 616 text=True, 617 timeout=5, 618 ) 619 total_system_bytes = int(result.stdout.strip()) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /scripts/check_gpu.py line 137
- Likely real
- Probably noise, 12% confidence
Starting a process with a partial executable path
136 try: 137 result = subprocess.run(['rocm-smi'], capture_output=True, text=True, timeout=5) 138 if result.returncode == 0: -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /scripts/check_gpu.py line 174
- Likely real
- Probably noise, 12% confidence
Starting a process with a partial executable path
173 try: 174 result = subprocess.run(['nvidia-smi'], capture_output=True, text=True, timeout=5) 175 if result.returncode == 0: -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 4% confidence
Loose version constraint: 'python-multipart>=0.0.18' — attacker could push malicious update
python-multipart>=0.0.18 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 33
- Likely real
- Probably noise, 4% confidence
Loose version constraint: 'python-multipart>=0.0.18' — attacker could push malicious update
python-multipart>=0.0.18 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /scripts/check_gpu.py line 174
- Likely real
- Probably noise, 3% confidence
subprocess call - check for execution of untrusted input.
173 try: 174 result = subprocess.run(['nvidia-smi'], capture_output=True, text=True, timeout=5) 175 if result.returncode == 0: -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/api/runtime_helpers.py line 62
- Likely real
- Probably noise, 3% confidence
subprocess call - check for execution of untrusted input.
61 62 app.state.tensorboard_process = subprocess.Popen( 63 [ 64 tensorboard_cmd, 65 "--logdir", 66 logdir, 67 "--port", 68 str(tensorboard_port), 69 "--bind_all", 70 ], 71 stdout=subprocess.DEVNULL, 72 stderr=subprocess.DEVNULL, 73 ) 74 return f"http://localhost:{tensorboard_port}" -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/audio_utils.py line 182
- Likely real
- Probably noise, 3% confidence
subprocess call - check for execution of untrusted input.
181 ] 182 subprocess.run(cmd, check=True, capture_output=True, timeout=120) 183 logger.debug(f"[AudioSaver] Saved audio to {output_path} (mp3, {target_sample_rate}Hz, {bitrate})") -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/gpu_config.py line 613
- Likely real
- Probably noise, 3% confidence
subprocess call - check for execution of untrusted input.
612 613 result = subprocess.run( 614 ["sysctl", "-n", "hw.memsize"], 615 capture_output=True, 616 text=True, 617 timeout=5, 618 ) 619 total_system_bytes = int(result.stdout.strip()) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/text_tasks/passphrase_store.py line 83
- Likely real
- Probably noise, 3% confidence
subprocess call - check for execution of untrusted input.
82 try: 83 result = subprocess.run( 84 [tool_path, "lookup", "service", service, "username", username], 85 stdout=subprocess.PIPE, 86 stderr=subprocess.PIPE, 87 text=True, 88 check=False, 89 timeout=_SECRET_TOOL_TIMEOUT_SEC, 90 ) 91 except subprocess.TimeoutExpired: -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/text_tasks/passphrase_store.py line 117
- Likely real
- Probably noise, 3% confidence
subprocess call - check for execution of untrusted input.
116 try: 117 result = subprocess.run( 118 [ 119 tool_path, 120 "store", 121 "--label", 122 "ACE-Step external LM passphrase", 123 "service", 124 service, 125 "username", 126 username, 127 ], 128 input=passphrase, 129 stdout=subprocess.PIPE, 130 stderr=subprocess.PIPE, 131 text=True, 132 check=False, 133 timeout=_SECRET_TOOL_TIMEOUT_SEC, 134 ) 135 except subprocess.TimeoutExpired: -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/text_tasks/secure_secret_store.py line 142
- Likely real
- Probably noise, 3% confidence
subprocess call - check for execution of untrusted input.
141 try: 142 return subprocess.run( 143 cmd, 144 input=stdin_bytes, 145 stdout=subprocess.PIPE, 146 stderr=subprocess.PIPE, 147 check=False, 148 timeout=_OPENSSL_TIMEOUT_SEC, 149 ) 150 except subprocess.TimeoutExpired as exc: -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /scripts/check_gpu.py line 137
- Likely real
- Probably noise, 3% confidence
subprocess call - check for execution of untrusted input.
136 try: 137 result = subprocess.run(['rocm-smi'], capture_output=True, text=True, timeout=5) 138 if result.returncode == 0: -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'torch>=2.9.1' — attacker could push malicious update
torch>=2.9.1; sys_platform == 'darwin' and platform_machine == 'arm64' -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'mlx>=0.25.2' — attacker could push malicious update
mlx>=0.25.2; sys_platform == 'darwin' and platform_machine == 'arm64' -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'mlx-lm>=0.20.0' — attacker could push malicious update
mlx-lm>=0.20.0; sys_platform == 'darwin' and platform_machine == 'arm64' -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'torchao>=0.16.0,<0.17.0' — attacker could push malicious update
torchao>=0.16.0,<0.17.0; platform_machine != 'aarch64' -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 41
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'torchao>=0.16.0,<0.17.0; platform_machine != 'aarch64'' — attacker could push malicious update
torchao>=0.16.0,<0.17.0; platform_machine != 'aarch64' -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'torchaudio>=2.9.1' — attacker could push malicious update
torchaudio>=2.9.1; sys_platform == 'darwin' and platform_machine == 'arm64' -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'vector-quantize-pytorch>=1.27.15' — attacker could push malicious update
vector-quantize-pytorch>=1.27.15 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'scipy>=1.10.1' — attacker could push malicious update
scipy>=1.10.1 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'soundfile>=0.13.1' — attacker could push malicious update
soundfile>=0.13.1 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'loguru>=0.7.3' — attacker could push malicious update
loguru>=0.7.3 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'einops>=0.8.1' — attacker could push malicious update
einops>=0.8.1 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'accelerate>=1.12.0' — attacker could push malicious update
accelerate>=1.12.0 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'fastapi>=0.110.0' — attacker could push malicious update
fastapi>=0.110.0 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'uvicorn>=0.27.0' — attacker could push malicious update
uvicorn[standard]>=0.27.0 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'numba>=0.63.1' — attacker could push malicious update
numba>=0.63.1 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'transformers>=4.51.0,<4.58.0' — attacker could push malicious update
transformers>=4.51.0,<4.58.0 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'torchcodec>=0.9.1' — attacker could push malicious update
torchcodec>=0.9.1; platform_machine != 'aarch64' -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'peft>=0.18.0' — attacker could push malicious update
peft>=0.18.0 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'lightning>=2.0.0' — attacker could push malicious update
lightning>=2.0.0 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'tensorboard>=2.0.0' — attacker could push malicious update
tensorboard>=2.0.0 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'tensorboard>=2.20.0' — attacker could push malicious update
tensorboard>=2.20.0 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'typer-slim>=0.21.1' — attacker could push malicious update
typer-slim>=0.21.1 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'pytorch-wavelets>=1.3.0' — attacker could push malicious update
pytorch-wavelets>=1.3.0 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'pywavelets>=1.9.0' — attacker could push malicious update
pywavelets>=1.9.0 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 22
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'transformers>=4.51.0,<4.58.0' — attacker could push malicious update
transformers>=4.51.0,<4.58.0 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'diffusers>=0.37.0' — attacker could push malicious update
diffusers>=0.37.0 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- package.json
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'vitepress^1.6.3' — attacker could push malicious update
vitepress:^1.6.3 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 50
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'tensorboard>=2.0.0' — attacker could push malicious update
tensorboard>=2.0.0 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 49
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'lightning>=2.0.0' — attacker could push malicious update
lightning>=2.0.0 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 47
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'peft>=0.18.0' — attacker could push malicious update
peft>=0.18.0 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 36
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'torchcodec>=0.9.1; platform_machine != 'aarch64'' — attacker could push malicious update
torchcodec>=0.9.1; platform_machine != 'aarch64' -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 35
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'vector-quantize-pytorch>=1.27.15' — attacker could push malicious update
vector-quantize-pytorch>=1.27.15 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 34
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'numba>=0.63.1' — attacker could push malicious update
numba>=0.63.1 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 31
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'fastapi>=0.110.0' — attacker could push malicious update
fastapi>=0.110.0 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 30
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'accelerate>=1.12.0' — attacker could push malicious update
accelerate>=1.12.0 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 29
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'einops>=0.8.1' — attacker could push malicious update
einops>=0.8.1 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 28
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'loguru>=0.7.3' — attacker could push malicious update
loguru>=0.7.3 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 27
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'soundfile>=0.13.1' — attacker could push malicious update
soundfile>=0.13.1 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 26
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'scipy>=1.10.1' — attacker could push malicious update
scipy>=1.10.1 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 25
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'matplotlib>=3.7.5' — attacker could push malicious update
matplotlib>=3.7.5 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- pyproject.toml
- Likely real
- Probably noise, 2% confidence
Loose version constraint: 'matplotlib>=3.7.5' — attacker could push malicious update
matplotlib>=3.7.5 -
- Engine
- Dependency checker · CVEs, typosquats and risky installs
- Counts towards
- Supply chain
- File
- requirements.txt line 32
- Likely real
- Probably noise, 1% confidence
Loose version constraint: 'uvicorn[standard]>=0.27.0' — attacker could push malicious update
uvicorn[standard]>=0.27.0 -
- Engine
- YARA · Malware signature rules
- Counts towards
- Malicious pattern
- File
- /install_uv.sh
- Likely real
- Probably noise, 0% confidence
YARA rule matched: detect_downloader — Downloads a remote resource - common and low-signal alone; see detect_download_and_execute for the composite version
-
- Engine
- YARA · Malware signature rules
- Counts towards
- Malicious pattern
- File
- /Dockerfile.jetson
- Likely real
- Probably noise, 0% confidence
YARA rule matched: detect_downloader — Downloads a remote resource - common and low-signal alone; see detect_download_and_execute for the composite version
-
- Engine
- YARA · Malware signature rules
- Counts towards
- Malicious pattern
- File
- /scripts/lora_data_prepare/gemini_caption.py
- Likely real
- Probably noise, 0% confidence
YARA rule matched: detect_downloader — Downloads a remote resource - common and low-signal alone; see detect_download_and_execute for the composite version
-
- Engine
- YARA · Malware signature rules
- Counts towards
- Malicious pattern
- File
- /Dockerfile
- Likely real
- Probably noise, 0% confidence
YARA rule matched: detect_downloader — Downloads a remote resource - common and low-signal alone; see detect_download_and_execute for the composite version
-
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py line 621
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
620 621 assert input_ids is None, "Only `input_ids` is supported for the lyric encoder." 622 assert attention_mask is not None, "Attention mask must be provided for the lyric encoder." -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/engine/sequence.py line 79
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
78 def block(self, i): 79 assert 0 <= i < self.num_blocks 80 return self.token_ids[i*self.block_size: (i+1)*self.block_size] -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/engine/model_runner.py line 262
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
261 def write_shm(self, method_name, *args): 262 assert self.world_size > 1 and self.rank == 0 263 data = _encode_for_shm([method_name, *args]) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/engine/model_runner.py line 254
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
253 def read_shm(self): 254 assert self.world_size > 1 and self.rank > 0 255 self.event.wait() -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/engine/block_manager.py line 146
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
145 else: 146 assert last_block.hash == -1 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/engine/block_manager.py line 139
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
138 elif len(seq) % self.block_size == 0: 139 assert last_block.hash == -1 140 token_ids = seq.block(seq.num_blocks-1) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/engine/block_manager.py line 134
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
133 if len(seq) % self.block_size == 1: 134 assert last_block.hash != -1 135 block_id = self.free_block_ids[0] -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/engine/block_manager.py line 81
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
80 f"free_blocks={len(self.free_block_ids)}") 81 assert not seq.block_table 82 h = -1 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/engine/block_manager.py line 71
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
70 def _deallocate_block(self, block_id: int) -> Block: 71 assert self.blocks[block_id].ref_count == 0 72 self.used_block_ids.remove(block_id) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/engine/block_manager.py line 64
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
63 block = self.blocks[block_id] 64 assert block.ref_count == 0 65 block.reset() -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/config.py line 26
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
25 self.max_model_len = min(self.max_model_len, self.hf_config.max_position_embeddings) 26 assert self.max_num_batched_tokens >= self.max_model_len -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/config.py line 23
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
22 assert self.kvcache_block_size % 256 == 0 23 assert 1 <= self.tensor_parallel_size <= 8 24 self.hf_config = AutoConfig.from_pretrained(self.model) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/config.py line 22
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
21 assert os.path.isdir(self.model) 22 assert self.kvcache_block_size % 256 == 0 23 assert 1 <= self.tensor_parallel_size <= 8 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/config.py line 21
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
20 def __post_init__(self): 21 assert os.path.isdir(self.model) 22 assert self.kvcache_block_size % 256 == 0 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py line 623
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
622 assert attention_mask is not None, "Attention mask must be provided for the lyric encoder." 623 assert inputs_embeds is not None, "Inputs embeddings must be provided for the lyric encoder." 624 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py line 622
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
621 assert input_ids is None, "Only `input_ids` is supported for the lyric encoder." 622 assert attention_mask is not None, "Attention mask must be provided for the lyric encoder." 623 assert inputs_embeds is not None, "Inputs embeddings must be provided for the lyric encoder." -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/layers/attention.py line 111
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
110 D = num_heads * head_dim 111 assert key.stride(-1) == 1 and value.stride(-1) == 1 112 assert key.stride(1) == head_dim and value.stride(1) == head_dim -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/models/xl_sft/modeling_acestep_v15_xl_base.py line 626
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
625 assert attention_mask is not None, "Attention mask must be provided for the lyric encoder." 626 assert inputs_embeds is not None, "Inputs embeddings must be provided for the lyric encoder." 627 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/models/xl_sft/modeling_acestep_v15_xl_base.py line 625
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
624 assert input_ids is None, "Only `input_ids` is supported for the lyric encoder." 625 assert attention_mask is not None, "Attention mask must be provided for the lyric encoder." 626 assert inputs_embeds is not None, "Inputs embeddings must be provided for the lyric encoder." -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/models/xl_sft/modeling_acestep_v15_xl_base.py line 624
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
623 624 assert input_ids is None, "Only `input_ids` is supported for the lyric encoder." 625 assert attention_mask is not None, "Attention mask must be provided for the lyric encoder." -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/models/xl_base/modeling_acestep_v15_xl_base.py line 626
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
625 assert attention_mask is not None, "Attention mask must be provided for the lyric encoder." 626 assert inputs_embeds is not None, "Inputs embeddings must be provided for the lyric encoder." 627 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/models/xl_base/modeling_acestep_v15_xl_base.py line 625
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
624 assert input_ids is None, "Only `input_ids` is supported for the lyric encoder." 625 assert attention_mask is not None, "Attention mask must be provided for the lyric encoder." 626 assert inputs_embeds is not None, "Inputs embeddings must be provided for the lyric encoder." -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/models/xl_base/modeling_acestep_v15_xl_base.py line 624
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
623 624 assert input_ids is None, "Only `input_ids` is supported for the lyric encoder." 625 assert attention_mask is not None, "Attention mask must be provided for the lyric encoder." -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/models/turbo/modeling_acestep_v15_turbo.py line 623
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
622 assert attention_mask is not None, "Attention mask must be provided for the lyric encoder." 623 assert inputs_embeds is not None, "Inputs embeddings must be provided for the lyric encoder." 624 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/models/turbo/modeling_acestep_v15_turbo.py line 622
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
621 assert input_ids is None, "Only `input_ids` is supported for the lyric encoder." 622 assert attention_mask is not None, "Attention mask must be provided for the lyric encoder." 623 assert inputs_embeds is not None, "Inputs embeddings must be provided for the lyric encoder." -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/models/turbo/modeling_acestep_v15_turbo.py line 621
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
620 621 assert input_ids is None, "Only `input_ids` is supported for the lyric encoder." 622 assert attention_mask is not None, "Attention mask must be provided for the lyric encoder." -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/models/sft/modeling_acestep_v15_base.py line 626
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
625 assert attention_mask is not None, "Attention mask must be provided for the lyric encoder." 626 assert inputs_embeds is not None, "Inputs embeddings must be provided for the lyric encoder." 627 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/models/sft/modeling_acestep_v15_base.py line 625
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
624 assert input_ids is None, "Only `input_ids` is supported for the lyric encoder." 625 assert attention_mask is not None, "Attention mask must be provided for the lyric encoder." 626 assert inputs_embeds is not None, "Inputs embeddings must be provided for the lyric encoder." -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/models/sft/modeling_acestep_v15_base.py line 624
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
623 624 assert input_ids is None, "Only `input_ids` is supported for the lyric encoder." 625 assert attention_mask is not None, "Attention mask must be provided for the lyric encoder." -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/models/base/modeling_acestep_v15_base.py line 626
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
625 assert attention_mask is not None, "Attention mask must be provided for the lyric encoder." 626 assert inputs_embeds is not None, "Inputs embeddings must be provided for the lyric encoder." 627 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/models/base/modeling_acestep_v15_base.py line 625
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
624 assert input_ids is None, "Only `input_ids` is supported for the lyric encoder." 625 assert attention_mask is not None, "Attention mask must be provided for the lyric encoder." 626 assert inputs_embeds is not None, "Inputs embeddings must be provided for the lyric encoder." -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/models/base/modeling_acestep_v15_base.py line 624
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
623 624 assert input_ids is None, "Only `input_ids` is supported for the lyric encoder." 625 assert attention_mask is not None, "Attention mask must be provided for the lyric encoder." -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/sampling_params.py line 27
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
26 if self.top_p is not None: 27 assert 0.0 < self.top_p <= 1.0, "top_p must be in (0.0, 1.0]" 28 assert self.repetition_penalty > 0.0, "repetition_penalty must be > 0.0" -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /scripts/flow_edit_overlay_smoke.py line 59
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
58 TGT_LYRICS = ex["lyrics"].replace(ORIG_VERSE1, NEW_VERSE1) 59 assert TGT_LYRICS != ex["lyrics"], "verse-1 replace marker not found" 60 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /scripts/flow_edit_overlay_smoke.py line 44
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
43 src_path = OUT / f"baseline_01_seed{SEED}.wav" 44 assert src_path.exists(), f"missing baseline wav at {src_path}; run flow_edit_smoke_test.py first" 45 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/training_v2/ui/summary.py line 97
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
96 97 assert console is not None 98 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/training_v2/ui/progress.py line 336
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
335 336 assert console is not None 337 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/training_v2/ui/errors.py line 180
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
179 180 assert console is not None 181 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/training_v2/ui/config_panel.py line 203
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
202 203 assert console is not None 204 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/training_v2/trainer_helpers.py line 312
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
311 module = trainer.module 312 assert module is not None 313 ckpt_dir = Path(resume_path) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/training_v2/trainer_helpers.py line 153
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
152 module = trainer.module 153 assert module is not None 154 os.makedirs(output_dir, exist_ok=True) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/training_v2/trainer_fixed.py line 226
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
225 cfg = self.training_config 226 assert self.module is not None 227 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/training_v2/trainer_basic_loop.py line 102
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
101 module = trainer.module 102 assert module is not None 103 -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/training_v2/preprocess_vae.py line 125
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
124 # Trim to actual written length 125 assert final_latents is not None 126 final_latents = final_latents[:, :, :latent_write_pos] -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/training_v2/preprocess_vae.py line 118
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
117 core_len = latent_core.shape[-1] 118 assert final_latents is not None 119 final_latents[:, :, latent_write_pos:latent_write_pos + core_len] = latent_core.cpu() -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/training/dataset_builder_modules/preprocess_vae.py line 146
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
145 # Trim to actual written length 146 assert final_latents is not None 147 final_latents = final_latents[:, :, :latent_write_pos] -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/training/dataset_builder_modules/preprocess_vae.py line 137
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
136 core_len = latent_core.shape[-1] 137 assert final_latents is not None 138 final_latents[:, :, latent_write_pos:latent_write_pos + core_len] = ( -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/sampling_params.py line 28
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
27 assert 0.0 < self.top_p <= 1.0, "top_p must be in (0.0, 1.0]" 28 assert self.repetition_penalty > 0.0, "repetition_penalty must be > 0.0" -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /scripts/flow_edit_overlay_smoke.py line 69
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
68 ) 69 assert ok 70 llm = LLMHandler() -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/sampling_params.py line 25
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
24 if self.top_k is not None: 25 assert self.top_k > 0, "top_k must be > 0" 26 if self.top_p is not None: -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/sampling_params.py line 23
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
22 assert self.temperature > 1e-10, "greedy sampling is not permitted" 23 assert self.cfg_scale >= 1.0, "cfg_scale must be >= 1.0" 24 if self.top_k is not None: -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/sampling_params.py line 22
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
21 def __post_init__(self): 22 assert self.temperature > 1e-10, "greedy sampling is not permitted" 23 assert self.cfg_scale >= 1.0, "cfg_scale must be >= 1.0" -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/models/qwen3.py line 110
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
109 ) 110 assert hidden_act == "silu" 111 self.act_fn = SiluAndMul() -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/models/qwen3.py line 35
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
34 self.total_num_kv_heads = num_kv_heads 35 assert self.total_num_kv_heads % tp_size == 0 36 self.num_kv_heads = self.total_num_kv_heads // tp_size -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/models/qwen3.py line 32
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
31 self.total_num_heads = num_heads 32 assert self.total_num_heads % tp_size == 0 33 self.num_heads = self.total_num_heads // tp_size -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/layers/rotary_embedding.py line 61
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
60 ): 61 assert rope_scaling is None 62 rotary_emb = RotaryEmbedding(head_size, rotary_dim, max_position, base) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/layers/rotary_embedding.py line 30
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
29 self.head_size = head_size 30 assert rotary_dim == head_size 31 inv_freq = 1.0 / (base**(torch.arange(0, rotary_dim, 2, dtype=torch.float) / rotary_dim)) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/layers/linear.py line 118
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
117 param_data = param.data 118 assert loaded_shard_id in ["q", "k", "v"] 119 if loaded_shard_id == "q": -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/layers/linear.py line 10
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
9 def divide(numerator, denominator): 10 assert numerator % denominator == 0 11 return numerator // denominator -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/layers/embed_head.py line 54
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
53 ): 54 assert not bias 55 super().__init__(num_embeddings, embedding_dim) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/layers/embed_head.py line 20
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
19 self.tp_size = dist_utils.get_world_size() 20 assert num_embeddings % self.tp_size == 0 21 self.num_embeddings = num_embeddings -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/layers/attention.py line 114
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
113 assert k_cache.stride(1) == D and v_cache.stride(1) == D 114 assert slot_mapping.numel() == N 115 store_kvcache_kernel[(N,)](key, key.stride(0), value, value.stride(0), k_cache, v_cache, slot_mapping, D) -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/layers/attention.py line 113
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
112 assert key.stride(1) == head_dim and value.stride(1) == head_dim 113 assert k_cache.stride(1) == D and v_cache.stride(1) == D 114 assert slot_mapping.numel() == N -
- Engine
- Bandit · Python security linter
- Counts towards
- Vulnerability
- File
- /acestep/third_parts/nano-vllm/nanovllm/layers/attention.py line 112
- Likely real
- Probably noise, 0% confidence
Use of assert detected. The enclosed code will be removed when compiling to optimised byte code.
111 assert key.stride(-1) == 1 and value.stride(-1) == 1 112 assert key.stride(1) == head_dim and value.stride(1) == head_dim 113 assert k_cache.stride(1) == D and v_cache.stride(1) == D
No findings match these filters.