Upload folder using huggingface_hub

README.md

@@ -0,0 +1,110 @@
+---
+license: mit
+pipeline_tag: text-classification
+tags:
+  - shell-safety
+  - classifier
+  - aprender
+  - rust
+  - bashrs
+model-index:
+  - name: paiml/shell-safety-classifier
+    results:
+      - task:
+          type: text-classification
+        dataset:
+          name: bashrs-corpus
+          type: custom
+        metrics:
+          - name: Train Accuracy
+            type: accuracy
+            value: 0.966
+          - name: Validation Accuracy
+            type: accuracy
+            value: 0.632
+          - name: Training Samples
+            type: custom
+            value: "17942"
+---
+
+# Shell Safety Classifier
+
+Classifies shell scripts into 5 safety categories using a lightweight MLP trained on the [bashrs](https://github.com/paiml/bashrs) corpus.
+
+## Labels
+
+| Index | Label | Description |
+|-------|-------|-------------|
+| 0 | safe | Script is deterministic, idempotent, and properly quoted |
+| 1 | needs-quoting | Contains unquoted variables susceptible to word splitting |
+| 2 | non-deterministic | Uses `$RANDOM`, timestamps, process IDs, or other non-deterministic sources |
+| 3 | non-idempotent | Operations not safe to re-run (missing `-p`, `-f` flags) |
+| 4 | unsafe | Security issues (injection vectors, privilege escalation) |
+
+## Architecture
+
+- **Model**: MLP classifier (ShellVocabulary token embeddings -> 128 -> 64 -> 5)
+- **Tokenizer**: ShellVocabulary (250 shell-specific tokens, max_seq_len=64)
+- **Format**: SafeTensors (model.safetensors) + JSON config + vocab
+- **Framework**: [aprender](https://github.com/paiml/aprender) (pure Rust ML, no Python dependencies)
+
+## Training
+
+- **Corpus**: bashrs v2 corpus (17,942 entries: 16,431 Bash + 804 Makefile + 707 Dockerfile)
+- **Split**: 80/20 train/validation (14,353 / 3,589)
+- **Epochs**: 50
+- **Optimizer**: Adam (lr=0.01)
+- **Loss**: CrossEntropyLoss
+- **Train accuracy**: 96.6%
+- **Validation accuracy**: 63.2%
+
+### Class Distribution
+
+| Label | Count | Percentage |
+|-------|-------|------------|
+| safe | 16,126 | 89.9% |
+| needs-quoting | 1,814 | 10.1% |
+| unsafe | 2 | 0.01% |
+
+## Usage
+
+### With bashrs CLI
+
+```bash
+# Classify a single script
+bashrs classify script.sh
+
+# Classify with format detection
+bashrs classify Makefile --format makefile
+
+# Multi-label classification
+bashrs classify script.sh --multi-label
+```
+
+### With aprender (Rust)
+
+```rust
+use aprender::models::shell_safety::{ShellSafetyClassifier, SafetyClass};
+
+let classifier = ShellSafetyClassifier::load("/path/to/model")?;
+let result = classifier.predict("echo $HOME")?;
+// result: SafetyClass::NeedsQuoting
+```
+
+## Files
+
+| File | Size | Description |
+|------|------|-------------|
+| model.safetensors | 68 KB | Model weights |
+| vocab.json | 3.6 KB | Shell tokenizer vocabulary |
+| config.json | 371 B | Model architecture config |
+
+## Limitations
+
+- The v2.0 MLP architecture has limited validation accuracy (63.2%) due to class imbalance and simple architecture
+- Best suited for binary safe/unsafe classification (96%+ accuracy when collapsing to 2 classes)
+- A Qwen2.5-Coder fine-tuned version is planned for higher accuracy on minority classes
+
+## License
+
+MIT

config.json

@@ -0,0 +1,19 @@
+{
+  "embed_dim": 64,
+  "framework": "aprender",
+  "hidden_dim": 128,
+  "labels": [
+    "safe",
+    "needs-quoting",
+    "non-deterministic",
+    "non-idempotent",
+    "unsafe"
+  ],
+  "max_seq_len": 64,
+  "model_type": "shell-safety-classifier",
+  "num_classes": 5,
+  "training_data": "bashrs-corpus",
+  "training_samples": 17942,
+  "version": "2.0",
+  "vocab_size": 251
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f8ae6168ed65ebdbcb442e4a7f854f0785bb737530d61755d5cadd8945f14188
+size 67991

vocab.json

@@ -0,0 +1,252 @@
+{
+  "$MACHTYPE": 170,
+  "(": 113,
+  "#!/usr/bin/env": 7,
+  "let": 42,
+  "-e": 176,
+  "$2": 163,
+  "cat": 53,
+  "=": 128,
+  "ln": 50,
+  "type": 26,
+  "home": 231,
+  "check": 252,
+  "{": 121,
+  "ls": 75,
+  "100": 211,
+  "||": 99,
+  "dest": 228,
+  "'": 200,
+  "build": 257,
+  "date": 81,
+  "head": 61,
+  "read": 10,
+  "-gt": 134,
+  "-v": 178,
+  "info": 246,
+  "then": 85,
+  "$HOME": 148,
+  "))": 116,
+  "sed": 55,
+  "return": 20,
+  "[EOS]": 4,
+  "unzip": 73,
+  "readonly": 41,
+  "local": 38,
+  "0": 204,
+  "255": 212,
+  "-k": 189,
+  "fi": 88,
+  "~": 147,
+  "$#": 158,
+  "<": 106,
+  "tee": 65,
+  "export": 13,
+  "if": 84,
+  "$HOSTNAME": 168,
+  "uniq": 59,
+  "5": 209,
+  "bg": 34,
+  "dirname": 79,
+  "root": 229,
+  "]": 118,
+  "-d": 140,
+  "exec": 18,
+  "builtin": 44,
+  "$RANDOM": 153,
+  "2>>": 110,
+  "jobs": 33,
+  "-ne": 132,
+  "tail": 62,
+  "declare": 39,
+  "-le": 135,
+  ";;": 102,
+  "-ge": 136,
+  "in": 90,
+  "$*": 160,
+  "kill": 32,
+  "$0": 161,
+  "--verbose": 195,
+  "ok": 242,
+  "!": 146,
+  "awk": 56,
+  "-g": 187,
+  "stat": 76,
+  "10": 210,
+  "4": 208,
+  "#!/bin/bash": 5,
+  "[SEP]": 3,
+  "$(": 125,
+  "test": 22,
+  "source": 16,
+  "typeset": 40,
+  "|": 98,
+  "$OSTYPE": 169,
+  "-a": 184,
+  "}": 122,
+  "output": 224,
+  "--no-clobber": 198,
+  "log": 220,
+  "path": 217,
+  "--quiet": 196,
+  "data": 226,
+  "etc": 234,
+  "-b": 185,
+  "gunzip": 71,
+  "exit": 19,
+  ";": 101,
+  "var": 235,
+  "debug": 247,
+  "2>&1": 111,
+  "grep": 54,
+  "-z": 137,
+  "$@": 159,
+  "err": 221,
+  "cd": 11,
+  "$$": 155,
+  "stop": 249,
+  "wc": 60,
+  "restart": 250,
+  "tar": 69,
+  "mkdir": 46,
+  "else": 86,
+  "user": 230,
+  "-h": 188,
+  "&&": 100,
+  "-l": 190,
+  "wait": 31,
+  "done": 92,
+  "$IFS": 165,
+  "-i": 177,
+  "lib": 233,
+  "$TMPDIR": 164,
+  "\\": 201,
+  "curl": 67,
+  "file": 215,
+  "$BASHPID": 154,
+  "src": 227,
+  "--force": 193,
+  "status": 251,
+  "-o": 183,
+  "$USER": 149,
+  "mktemp": 80,
+  "[[": 119,
+  "$": 123,
+  "[": 117,
+  "-u": 191,
+  "$((": 126,
+  "-w": 143,
+  "error": 243,
+  "[UNK]": 1,
+  "unset": 14,
+  "$1": 162,
+  "trap": 30,
+  "-rf": 174,
+  "config": 225,
+  "remove": 255,
+  "${": 124,
+  "-p": 171,
+  "-c": 186,
+  "755": 214,
+  "install": 253,
+  "$SHELL": 152,
+  "run": 240,
+  "do": 91,
+  "-x": 192,
+  "fatal": 245,
+  "#!/bin/sh": 6,
+  "cp": 48,
+  "--parents": 197,
+  "touch": 74,
+  "644": 213,
+  "-s": 181,
+  "gzip": 70,
+  "zip": 72,
+  "-f": 172,
+  "chmod": 51,
+  "$!": 156,
+  "mv": 49,
+  "umask": 36,
+  "--recursive": 194,
+  "for": 89,
+  "3": 207,
+  "input": 223,
+  "dev": 236,
+  "-n": 175,
+  "$?": 157,
+  "-q": 179,
+  "fg": 35,
+  ">>": 105,
+  "&>": 112,
+  "wget": 68,
+  "echo": 8,
+  "!=": 130,
+  "xargs": 66,
+  "1": 205,
+  "]]": 120,
+  "rm": 47,
+  "start": 248,
+  "sleep": 82,
+  "name": 218,
+  "sys": 239,
+  "case": 95,
+  "eval": 17,
+  "printf": 9,
+  "until": 94,
+  "-m": 180,
+  "find": 57,
+  "command": 43,
+  ")": 114,
+  "dir": 216,
+  "getopts": 37,
+  "function": 97,
+  "&": 103,
+  "==": 129,
+  ":": 25,
+  "enable": 45,
+  "while": 93,
+  "hash": 27,
+  "sort": 58,
+  "\"": 199,
+  "tmp": 219,
+  "$PATH": 150,
+  "update": 254,
+  "2>": 109,
+  "esac": 96,
+  "shift": 21,
+  "-r": 173,
+  "\t": 203,
+  "[CLS]": 2,
+  "<<": 107,
+  "false": 24,
+  "`": 127,
+  "\n": 202,
+  "[PAD]": 0,
+  "alias": 28,
+  "true": 23,
+  "tr": 64,
+  "bin": 232,
+  "pwd": 12,
+  "null": 237,
+  "$TERM": 167,
+  "cut": 63,
+  "basename": 78,
+  ">": 104,
+  "proc": 238,
+  "((": 115,
+  "<<<": 108,
+  "$LANG": 166,
+  "2": 206,
+  "chown": 52,
+  "-lt": 133,
+  "-t": 182,
+  "out": 222,
+  "-eq": 131,
+  "set": 15,
+  "$PWD": 151,
+  "elif": 87,
+  "opt": 241,
+  "clean": 256,
+  "unalias": 29,
+  "warning": 244
+}