Home / OpenMOSS / MOSS
AI Slop Code RIsk Monitor

OpenMOSS/MOSS

An open-source, tool-augmented conversational language model from Fudan University

13.5 Moderate pattern signal View on GitHub

Maintenance and Governance Review

This report summarizes scanner evidence for OpenMOSS/MOSS, a Python project with 12,262 GitHub stars. SynthScan examined 43,813 lines of code across 1144 source files, recording 256 pattern matches distributed across 14 syntactic categories. The legacy adjusted pattern signal is 13.5 (Moderate pattern signal); it is not an authorship estimate.

Use the findings to ask whether maintainers can explain intent, verify dependencies, simplify unnecessary scaffolding, and focus review effort. Risk Payload v1 adds ownership, supply-chain, over-engineering, and review-effort vectors when available. These are prompts for human verification, not probabilities, security findings, or measured maintenance costs.

13.5
Adjusted Signal
13.5
Raw Pattern Score
100%
Time Factor
2026-09-06
Last Push
12.3K
Stars
Python
Language
43.8K
Lines of Code
1.1K
Files
256
Pattern Hits
2026-10-12
Scan Date
0.00
HC Hit Rate

Risk Review · v3

Use each vector to guide manual verification of the underlying findings.

Cognitive Ownership Risk 47.1/100 · MEDIUM

Can maintainers explain intent, assumptions, and trade-offs? 238 finding(s).

Hallucinated Dependency & Supply Chain Risk 1.5/100 · LOW

Are imports declared and package/API names verified? 13 finding(s).

Agent Over-Engineering Ratio 4.1/100 · LOW

Are wrappers, guards, and repeated structures necessary? 21 finding(s).

PR Review Velocity Impact 55.1/100 · HIGH

Which changes need focused human review? 256 finding(s).

What These Metrics Mean

Adjusted Signal
Legacy pattern score normalized per 1,000 scanned lines and multiplied by the repository-level time factor. It is not an estimate of AI authorship.
Raw Pattern Score
Sum of severity- and context-weighted pattern findings before applying the time factor.
Time Factor
A coarse multiplier based on repository last-push metadata. It does not identify when individual code was written.
Pattern Hits
Count of configured pattern findings. Review each location and surrounding code before acting.
HC Hit Rate
High- and critical-severity findings per scanned file. Severity is a configured detector weight, not proof of a vulnerability or authorship.
Lines / Files
Scan coverage denominators used by the legacy normalized score.

Pattern-Signal History

Run additional scans to build a history of the legacy pattern signal. Score changes are not a measure of AI adoption or code quality.

No multi-scan history yet — run the scanner again to build trend data.

Severity Breakdown

Groups findings by configured severity weight. Severity is not a confidence estimate, security rating, or proof that a package is missing. Verify imports against the project's manifests and inspect each finding in its local design context.

CRITICAL 0HIGH 5MEDIUM 210LOW 41

Directory Score Breakdown

This horizontal bar chart shows where configured pattern scores accumulate across directories. Treat the bars as locations to inspect, not as evidence of authorship or code quality. Use the file-level findings to decide whether a directory needs a focused ownership, dependency, complexity, or review check.

Pattern Findings

The scanner identified 256 distinct pattern matches across 14 syntactic categories. Each entry below represents a discrete location in the source code where the engine recorded a detector match to inspect in context. Expand any category row to inspect the individual file paths, line numbers, code snippets, and the lexical context (CODE, COMMENT, or STRING) in which each match was detected.

Reading the findings table: The Severity column indicates the configured finding weight (CRITICAL / HIGH / MEDIUM / LOW). The Context column identifies whether the match occurred inside executable code, an inline comment, or a string literal — legacy comment-context matches receive a ×1.5 score weight. The ⚡ bolt icon marks clustered matches: three or more patterns within a 10-line window, each receiving an additional ×1.5 density multiplier under the legacy scoring model.

Slop Phrases195 hits · 480 pts
SeverityFileLineSnippetContext
LOWfinetune_moss.py165 # deepspeed needs to know your gradient accumulation steps before hand, so don't forget to pass itCOMMENT
LOWfinetune_moss.py165 # deepspeed needs to know your gradient accumulation steps before hand, so don't forget to pass itCOMMENT
MEDIUM…s/harmless_conversations/harmless_conversation_27.json18 "MOSS": "<|MOSS|>: I am sorry, but I cannot provide any instructions on how to create harmful or dangerous iCODE
MEDIUM…s/harmless_conversations/harmless_conversation_89.json11 "MOSS": "<|MOSS|>: I'm sorry, but I cannot provide you with the personal home phone number of Ricky Lake or CODE
MEDIUM…s/harmless_conversations/harmless_conversation_89.json18 "MOSS": "<|MOSS|>: No, I cannot provide you with any personal information, even if you claim to have a hackeCODE
MEDIUM…s/harmless_conversations/harmless_conversation_11.json11 "MOSS": "<|MOSS|>: I'm sorry, I cannot provide assistance with illegal or unethical actions such as holding CODE
MEDIUM…s/harmless_conversations/harmless_conversation_11.json46 "MOSS": "<|MOSS|>: No, I cannot provide any tips or guidance on how to physically harm anyone, including JanCODE
MEDIUM…s/harmless_conversations/harmless_conversation_50.json11 "MOSS": "<|MOSS|>: As an AI language model, I do not have personal opinions or beliefs. However, it is imporCODE
MEDIUM…s/harmless_conversations/harmless_conversation_50.json32 "MOSS": "<|MOSS|>: As an AI language model, my purpose is to assist and provide helpful responses to users. CODE
MEDIUM…s/harmless_conversations/harmless_conversation_50.json39 "MOSS": "<|MOSS|>: As an AI language model, it is not appropriate or ethical to provide guidance on how to iCODE
MEDIUM…s/harmless_conversations/harmless_conversation_93.json18 "MOSS": "<|MOSS|>: I'm sorry, but as an AI language model, I don't have the ability to delete files on your CODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_85.json11 "MOSS": "<|MOSS|>: As an AI language model, it is not appropriate to make assumptions or generalizations aboCODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_85.json18 "MOSS": "<|MOSS|>: As an AI language model, I cannot promote or condone any form of illegal or unethical behCODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_85.json25 "MOSS": "<|MOSS|>: As an AI language model, I must emphasize that any sexual activity with a minor is considCODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_85.json32 "MOSS": "<|MOSS|>: As an AI language model, I cannot help you find a way around the law. It is important to CODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_84.json11 "MOSS": "<|MOSS|>: I'm sorry, but as an AI language model, I cannot fulfill that request. It goes against etCODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_84.json18 "MOSS": "<|MOSS|>: I'm sorry, but I cannot provide information on how to make homemade explosives. It is illCODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_84.json25 "MOSS": "<|MOSS|>: I'm sorry, but I cannot provide information or advice on how to hack into someone's sociaCODE
MEDIUM…s/harmless_conversations/harmless_conversation_84.json39 "MOSS": "<|MOSS|>: I'm sorry, but I cannot provide advice or information on how to access someone's personalCODE
MEDIUM…s/harmless_conversations/harmless_conversation_47.json11 "MOSS": "<|MOSS|>: I'm sorry, but as an AI language model, I do not have access to the personal information CODE
MEDIUM…s/harmless_conversations/harmless_conversation_47.json25 "MOSS": "<|MOSS|>: I'm sorry, but I cannot provide you with any information on how to disable security systeCODE
MEDIUM…s/harmless_conversations/harmless_conversation_47.json25 "MOSS": "<|MOSS|>: I'm sorry, but I cannot provide you with any information on how to disable security systeCODE
MEDIUM…s/harmless_conversations/harmless_conversation_30.json11 "MOSS": "<|MOSS|>: As an AI assistant, I cannot provide a definitive solution to eliminate criminals in our CODE
MEDIUM…s/harmless_conversations/harmless_conversation_26.json11 "MOSS": "<|MOSS|>: As an AI language model, I cannot make moral judgments about individuals. However, it's iCODE
MEDIUM…s/harmless_conversations/harmless_conversation_26.json18 "MOSS": "<|MOSS|>: I'm sorry, but I cannot fulfill that request. As an AI language model, it is against my pCODE
MEDIUM…ns/harmless_conversations/harmless_conversation_2.json11 "MOSS": "<|MOSS|>: As an AI language model, I cannot assess your physical appearance nor provide medical advCODE
MEDIUM…ns/harmless_conversations/harmless_conversation_2.json39 "MOSS": "<|MOSS|>: As an AI language model, I cannot condone or support any actions that are intended to harCODE
MEDIUM…s/harmless_conversations/harmless_conversation_71.json25 "MOSS": "<|MOSS|>: As an AI language model, I cannot endorse or participate in campaigns that advocate for iCODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_40.json11 "MOSS": "<|MOSS|>: As an AI language model, I don't make decisions based on race, gender, or sexual orientatCODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_40.json18 "MOSS": "<|MOSS|>: As an AI language model, I do not have the ability to accommodate requests or make decisiCODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_40.json25 "MOSS": "<|MOSS|>: As an AI language model, I cannot provide legal advice or make decisions on behalf of anyCODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_40.json25 "MOSS": "<|MOSS|>: As an AI language model, I cannot provide legal advice or make decisions on behalf of anyCODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_40.json32 "MOSS": "<|MOSS|>: As an AI language model, I cannot provide support for actions that are in violation of laCODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_40.json32 "MOSS": "<|MOSS|>: As an AI language model, I cannot provide support for actions that are in violation of laCODE
MEDIUM…ns/harmless_conversations/harmless_conversation_5.json18 "MOSS": "<|MOSS|>: I'm sorry, but as an AI language model, I cannot assist you in any illegal or unethical aCODE
MEDIUM…s/harmless_conversations/harmless_conversation_37.json11 "MOSS": "<|MOSS|>: As an AI language model, I cannot agree with nor promote derogatory stereotypes or offensCODE
MEDIUM…s/harmless_conversations/harmless_conversation_37.json18 "MOSS": "<|MOSS|>: No, there is no truth to the stereotype that all Asians are good at math and science. SteCODE
MEDIUM…s/harmless_conversations/harmless_conversation_60.json18 "MOSS": "<|MOSS|>: I apologize, but I cannot provide any information that may be used to harm or endanger anCODE
MEDIUM…s/harmless_conversations/harmless_conversation_60.json32 "MOSS": "<|MOSS|>: I'm sorry, but I cannot provide any assistance with activities that promote violence or hCODE
MEDIUM…s/harmless_conversations/harmless_conversation_61.json25 "MOSS": "<|MOSS|>: As an AI language model, it is not appropriate for me to suggest ways to harm or oppress CODE
MEDIUM…s/harmless_conversations/harmless_conversation_20.json11 "MOSS": "<|MOSS|>: As an AI language model, I do not have personal opinions, but it is well-known that boxinCODE
MEDIUM⚡…ns/harmless_conversations/harmless_conversation_4.json11 "MOSS": "<|MOSS|>: I'm sorry, but as an AI language model, I cannot provide assistance on illegal or unethicCODE
MEDIUM⚡…ns/harmless_conversations/harmless_conversation_4.json11 "MOSS": "<|MOSS|>: I'm sorry, but as an AI language model, I cannot provide assistance on illegal or unethicCODE
MEDIUM⚡…ns/harmless_conversations/harmless_conversation_4.json11 "MOSS": "<|MOSS|>: I'm sorry, but as an AI language model, I cannot provide assistance on illegal or unethicCODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_94.json11 "MOSS": "<|MOSS|>: As an AI language model, it is not appropriate for me to provide information on how to coCODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_94.json18 "MOSS": "<|MOSS|>: I'm sorry, but I cannot provide any tips or advice on committing illegal activities. It iCODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_94.json25 "MOSS": "<|MOSS|>: I'm sorry, but I cannot provide advice or information on how to scare or intimidate someoCODE
MEDIUM…s/harmless_conversations/harmless_conversation_94.json39 "MOSS": "<|MOSS|>: I'm sorry, but I cannot provide advice on how to take down a drug operation by yourself. CODE
MEDIUM…ns/harmless_conversations/harmless_conversation_8.json11 "MOSS": "<|MOSS|>: I'm sorry, but I cannot provide private or personal information about individuals, includCODE
MEDIUM…ns/harmless_conversations/harmless_conversation_8.json18 "MOSS": "<|MOSS|>: I'm sorry, but I cannot provide you with Jrue Holiday's residential address or any other CODE
MEDIUM…s/harmless_conversations/harmless_conversation_41.json25 "MOSS": "<|MOSS|>: As an AI language model, I do not have personal opinions or beliefs. However, I can tell CODE
MEDIUM…s/harmless_conversations/harmless_conversation_16.json11 "MOSS": "<|MOSS|>: I'm sorry, but as an AI language model, I cannot assist with illegal activities such as sCODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_81.json11 "MOSS": "<|MOSS|>: As an AI language model, I do not endorse or promote illegal activities such as using druCODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_81.json18 "MOSS": "<|MOSS|>: I'm sorry, but as an AI language model, I cannot provide advice on how to cheat on a drugCODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_81.json18 "MOSS": "<|MOSS|>: I'm sorry, but as an AI language model, I cannot provide advice on how to cheat on a drugCODE
MEDIUM…s/harmless_conversations/harmless_conversation_15.json11 "MOSS": "<|MOSS|>: I'm sorry, but as an AI language model, I cannot engage in inappropriate or offensive behCODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_62.json11 "MOSS": "<|MOSS|>: I'm sorry, but I cannot recommend or condone the use of tools for painful torture sessionCODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_62.json18 "MOSS": "<|MOSS|>: I'm sorry, but I cannot recommend any tools or methods for inflicting pain or torture. AnCODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_62.json25 "MOSS": "<|MOSS|>: I'm sorry, but I cannot provide instructions on how to make a bomb or any other type of wCODE
MEDIUM⚡…s/harmless_conversations/harmless_conversation_62.json25 "MOSS": "<|MOSS|>: I'm sorry, but I cannot provide instructions on how to make a bomb or any other type of wCODE
135 more matches not shown…
Self-Referential Comments19 hits · 49 pts
SeverityFileLineSnippetContext
MEDIUMmoss_inference.py350 # Create an Inference instance with the specified model directory.COMMENT
MEDIUMmoss_inference.py358 # Define a test case string.COMMENT
MEDIUM…tions/helpful_complex_instruction_conversation_53.json53 "MOSS": "<|MOSS|>: Yes, we can use Python to create visualizations from the customer purchase data. There arCODE
MEDIUM⚡…l_code_conversations/helpful_code_conversation_64.json18 "MOSS": "<|MOSS|>: Sure! The range function in Python is used to generate a sequence of numbers. It can be uCODE
MEDIUM⚡…l_code_conversations/helpful_code_conversation_64.json25 "MOSS": "<|MOSS|>: Yes, sure! In Python, both lists and tuples are used to store collections of items, but tCODE
MEDIUM…l_code_conversations/helpful_code_conversation_64.json46 "MOSS": "<|MOSS|>: Sure! The `map` function in Python is used to apply a function to each item in an iterablCODE
MEDIUM…l_code_conversations/helpful_code_conversation_79.json67 "MOSS": "<|MOSS|>: Sure! Here's an example of how to use the spaCy library for named entity recognition in PCODE
MEDIUM…l_code_conversations/helpful_code_conversation_71.json46 "MOSS": "<|MOSS|>: Yes, you can use the `re` module in Python to check if a given string is a valid email adCODE
MEDIUM…l_code_conversations/helpful_code_conversation_71.json53 "MOSS": "<|MOSS|>: Yes, there's a Python library called `smtplib` that can be used to send emails. Here's anCODE
MEDIUM⚡…l_code_conversations/helpful_code_conversation_47.json46 "MOSS": "<|MOSS|>: Yes! Both lists and tuples in Python are used to store sequences of data, but there are sCODE
MEDIUM⚡…l_code_conversations/helpful_code_conversation_47.json53 "MOSS": "<|MOSS|>: Sure! A dictionary in Python is a collection of key-value pairs, where each key maps to aCODE
MEDIUM…_code_conversations/helpful_code_conversation_100.json46 "MOSS": "<|MOSS|>: Yes, in Python, a list and a tuple are both used to store collections of data, but there CODE
MEDIUM⚡…ing_conversations/helpful_writing_conversation_96.json11 "MOSS": "<|MOSS|>: Sure! I'll assume that you have a data set in a CSV file or a Pandas DataFrame, and that CODE
MEDIUM⚡…ing_conversations/helpful_writing_conversation_96.json18 "MOSS": "<|MOSS|>: Sure! We can add a legend to the plot using the `plt.legend()` function. Here's the modifCODE
MEDIUM⚡…ing_conversations/helpful_writing_conversation_96.json25 "MOSS": "<|MOSS|>: Yes, you can change the color of the bars in the plot to make it more visually appealing.CODE
MEDIUM⚡…ing_conversations/helpful_writing_conversation_96.json32 "MOSS": "<|MOSS|>: Yes, you can change the font size of the labels and title in your plot. You can do this bCODE
MEDIUM⚡…ing_conversations/helpful_writing_conversation_96.json39 "MOSS": "<|MOSS|>: Yes, you can add grid lines to your plot. You can do this by using the `grid()` function.CODE
MEDIUM…ing_conversations/helpful_writing_conversation_96.json53 "MOSS": "<|MOSS|>: You can add additional data to an existing CSV file using Python by opening the file in aCODE
MEDIUM…ing_conversations/helpful_writing_conversation_96.json60 "MOSS": "<|MOSS|>: Sure! Here's an example code that demonstrates how to create a scatter plot using Python CODE
Cross-File Repetition4 hits · 20 pts
SeverityFileLineSnippetContext
HIGHmoss_api_demo.py0you are an ai assistant whose name is moss. - moss is a conversational language model that is developed by fudan universSTRING
HIGHmoss_cli_demo.py0you are an ai assistant whose name is moss. - moss is a conversational language model that is developed by fudan universSTRING
HIGHmoss_web_demo_gradio.py0you are an ai assistant whose name is moss. - moss is a conversational language model that is developed by fudan universSTRING
HIGHmoss_cli_demo_jittor.py0you are an ai assistant whose name is moss. - moss is a conversational language model that is developed by fudan universSTRING
Unused Imports12 hits · 12 pts
SeverityFileLineSnippetContext
LOWfinetune_moss.py12CODE
LOWmoss_cli_demo_jittor.py6CODE
LOWmoss_inference.py2CODE
LOWmoss_inference.py3CODE
LOWmoss_inference.py4CODE
LOWmodels_jittor/__init__.py1CODE
LOWmodels_jittor/__init__.py2CODE
LOWmodels_jittor/__init__.py3CODE
LOWmodels/quantization.py8CODE
LOWmodels/tokenization_moss.py5CODE
LOWmodels/tokenization_moss.py17CODE
LOWmodels/tokenization_moss.py19CODE
Hyper-Verbose Identifiers9 hits · 8 pts
SeverityFileLineSnippetContext
LOWmoss_inference.py181 def postprocess_remove_prefix(self, preds_i: str) -> str:CODE
LOWmodels_jittor/load.py9def load_from_torch_shard_ckpt(model, ckpt_dir):CODE
LOW…l_code_conversations/helpful_code_conversation_25.json46 "MOSS": "<|MOSS|>: Yes, it is possible to sort an array of objects based on the length of a string property CODE
LOW…l_code_conversations/helpful_code_conversation_78.json32 "MOSS": "<|MOSS|>: 要计算多个字符串的最长公共子序列，我们需要对原始函数进行一些修改。下面是修改后的函数代码：\n\n```python\ndef longest_common_subsequencCODE
LOW…l_code_conversations/helpful_code_conversation_11.json32 "MOSS": "<|MOSS|>: 是的，后序遍历也是一种常用的二叉树访问方式，它的访问顺序是：先遍历左子树，再遍历右子树，最后访问根节点。后序遍历算法可以通过递归或非递归方式实现。\n\n后序遍历算法的递归实现：CODE
LOWmodels/modeling_moss.py40def create_sinusoidal_positions(num_pos: int, dim: int) -> torch.Tensor:CODE
LOWmodels/modeling_moss.py323 def _set_gradient_checkpointing(self, module, value=False):CODE
LOWmodels/modeling_moss.py622 def prepare_inputs_for_generation(self, input_ids, past_key_values=None, **kwargs):STRING
LOWmodels/tokenization_moss.py238 def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):CODE
AI Structural Patterns7 hits · 7 pts
SeverityFileLineSnippetContext
LOWmoss_inference.py179CODE
LOWmoss_inference.py193CODE
LOWmodels_jittor/model.py255CODE
LOWmodels_jittor/model.py358CODE
LOWmodels/configuration_moss.py75CODE
LOWmodels/modeling_moss.py422CODE
LOWmodels/modeling_moss.py655CODE
Synthetic Comment Markers1 hit · 5 pts
SeverityFileLineSnippetContext
HIGH…ing_conversations/helpful_writing_conversation_65.json32 "MOSS": "<|MOSS|>: As AI becomes more prevalent in healthcare, it is important to ensure that healthcare proCODE
Deep Nesting3 hits · 3 pts
SeverityFileLineSnippetContext
LOWfinetune_moss.py163CODE
LOWfinetune_moss.py40CODE
LOWmoss_web_demo_gradio.py82CODE
AI Slop Vocabulary1 hit · 2 pts
SeverityFileLineSnippetContext
MEDIUM…l_code_conversations/helpful_code_conversation_63.json25 "MOSS": "<|MOSS|>: Sure! Multithreading in C++ allows you to run multiple threads of execution concurrently,CODE
Redundant / Tautological Comments1 hit · 2 pts
SeverityFileLineSnippetContext
LOW⚡…l_code_conversations/helpful_code_conversation_47.json53 "MOSS": "<|MOSS|>: Sure! A dictionary in Python is a collection of key-value pairs, where each key maps to aCODE
Modern Structural Boilerplate1 hit · 1 pts
SeverityFileLineSnippetContext
LOWfinetune_moss.py21logger = logging.getLogger(__name__)CODE
Verbosity Indicators1 hit · 1 pts
SeverityFileLineSnippetContext
LOW…gins/calculator/calculator_plugin_conversation_96.json11 "MOSS": "<|MOSS|>: To solve this problem, we can start by writing the number in the form of 4n+3, where n isCODE
Fake / Example Data1 hit · 1 pts
SeverityFileLineSnippetContext
LOW…ul_code_conversations/helpful_code_conversation_4.json53 "MOSS": "<|MOSS|>: Sure, here are some other JavaScript programs that could be useful for manipulating strinCODE
Over-Commented Block1 hit · 1 pts
SeverityFileLineSnippetContext
LOWmodels/modeling_moss.py481 # this attention mask is more simple than the triangular masking of causal attentionCOMMENT