Instructions to use edoggy/audit-coder-7b-v1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use edoggy/audit-coder-7b-v1 with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf edoggy/audit-coder-7b-v1:Q4_K_M # Run inference directly in the terminal: llama cli -hf edoggy/audit-coder-7b-v1:Q4_K_M
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf edoggy/audit-coder-7b-v1:Q4_K_M # Run inference directly in the terminal: llama cli -hf edoggy/audit-coder-7b-v1:Q4_K_M
Build from source code
git clone https://github.com/ggml-org/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-app # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama serve -hf edoggy/audit-coder-7b-v1:Q4_K_M # Run inference directly in the terminal: ./build/bin/llama cli -hf edoggy/audit-coder-7b-v1:Q4_K_M
Use Docker
docker model run hf.co/edoggy/audit-coder-7b-v1:Q4_K_M
- Llama new
- LM Studio
- Jan
- Ollama
How to use edoggy/audit-coder-7b-v1 with Ollama:
ollama run hf.co/edoggy/audit-coder-7b-v1:Q4_K_M
- Unsloth Desktop
- Pi
How to use edoggy/audit-coder-7b-v1 with Pi:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf edoggy/audit-coder-7b-v1:Q4_K_M
Configure the model in Pi
# Install Pi: npm install -g @earendil-works/pi-coding-agent # Add to ~/.pi/agent/models.json: { "providers": { "llama-cpp": { "baseUrl": "http://localhost:8080/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "edoggy/audit-coder-7b-v1:Q4_K_M" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Docker Model Runner
How to use edoggy/audit-coder-7b-v1 with Docker Model Runner:
docker model run hf.co/edoggy/audit-coder-7b-v1:Q4_K_M
- Lemonade
How to use edoggy/audit-coder-7b-v1 with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull edoggy/audit-coder-7b-v1:Q4_K_M
Run and chat with the model
lemonade run user.audit-coder-7b-v1-Q4_K_M
List all available models
lemonade list
- Hermes Agent
How to use edoggy/audit-coder-7b-v1 with Hermes Agent:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf edoggy/audit-coder-7b-v1:Q4_K_M
Configure Hermes
# Install Hermes: curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash hermes setup # Point Hermes at the local server: hermes config set model.provider custom hermes config set model.base_url http://127.0.0.1:8080/v1 hermes config set model.default edoggy/audit-coder-7b-v1:Q4_K_M
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use edoggy/audit-coder-7b-v1 with OpenClaw:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf edoggy/audit-coder-7b-v1:Q4_K_M
Configure OpenClaw
# Install OpenClaw: npm install -g openclaw@latest # Register the local server and set it as the default model: openclaw onboard --non-interactive --mode local \ --auth-choice custom-api-key \ --custom-base-url http://127.0.0.1:8080/v1 \ --custom-model-id "edoggy/audit-coder-7b-v1:Q4_K_M" \ --custom-provider-id llama-cpp \ --custom-compatibility openai \ --custom-text-input \ --accept-risk \ --skip-health
Run OpenClaw
openclaw agent --local --agent main --message "Hello from Hugging Face"
Upload Modelfile with huggingface_hub
Browse files
Modelfile
ADDED
|
@@ -0,0 +1,95 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
FROM ./audit-7b-v1.Q4_K_M.gguf
|
| 2 |
+
TEMPLATE """{{- if .Suffix }}<|fim_prefix|>{{ .Prompt }}<|fim_suffix|>{{ .Suffix }}<|fim_middle|>
|
| 3 |
+
{{- else if .Messages }}
|
| 4 |
+
{{- if or .System .Tools }}<|im_start|>system
|
| 5 |
+
{{- if .System }}
|
| 6 |
+
{{ .System }}
|
| 7 |
+
{{- end }}
|
| 8 |
+
{{- if .Tools }}
|
| 9 |
+
|
| 10 |
+
# Tools
|
| 11 |
+
|
| 12 |
+
You may call one or more functions to assist with the user query.
|
| 13 |
+
|
| 14 |
+
You are provided with function signatures within <tools></tools>:
|
| 15 |
+
<tools>
|
| 16 |
+
{{- range .Tools }}
|
| 17 |
+
{"type": "function", "function": {{ .Function }}}
|
| 18 |
+
{{- end }}
|
| 19 |
+
</tools>
|
| 20 |
+
|
| 21 |
+
For each function call, return a json object with function name and arguments within <tool_call></tool_call> with NO other text. Do not include any backticks or ```json.
|
| 22 |
+
<tool_call>
|
| 23 |
+
{"name": <function-name>, "arguments": <args-json-object>}
|
| 24 |
+
</tool_call>
|
| 25 |
+
{{- end }}<|im_end|>
|
| 26 |
+
{{ end }}
|
| 27 |
+
{{- range $i, $_ := .Messages }}
|
| 28 |
+
{{- $last := eq (len (slice $.Messages $i)) 1 -}}
|
| 29 |
+
{{- if eq .Role "user" }}<|im_start|>user
|
| 30 |
+
{{ .Content }}<|im_end|>
|
| 31 |
+
{{ else if eq .Role "assistant" }}<|im_start|>assistant
|
| 32 |
+
{{ if .Content }}{{ .Content }}
|
| 33 |
+
{{- else if .ToolCalls }}<tool_call>
|
| 34 |
+
{{ range .ToolCalls }}{"name": "{{ .Function.Name }}", "arguments": {{ .Function.Arguments }}}
|
| 35 |
+
{{ end }}</tool_call>
|
| 36 |
+
{{- end }}{{ if not $last }}<|im_end|>
|
| 37 |
+
{{ end }}
|
| 38 |
+
{{- else if eq .Role "tool" }}<|im_start|>user
|
| 39 |
+
<tool_response>
|
| 40 |
+
{{ .Content }}
|
| 41 |
+
</tool_response><|im_end|>
|
| 42 |
+
{{ end }}
|
| 43 |
+
{{- if and (ne .Role "assistant") $last }}<|im_start|>assistant
|
| 44 |
+
{{ end }}
|
| 45 |
+
{{- end }}
|
| 46 |
+
{{- else }}
|
| 47 |
+
{{- if .System }}<|im_start|>system
|
| 48 |
+
{{ .System }}<|im_end|>
|
| 49 |
+
{{ end }}{{ if .Prompt }}<|im_start|>user
|
| 50 |
+
{{ .Prompt }}<|im_end|>
|
| 51 |
+
{{ end }}<|im_start|>assistant
|
| 52 |
+
{{ end }}{{ .Response }}{{ if .Response }}<|im_end|>{{ end }}"""
|
| 53 |
+
|
| 54 |
+
SYSTEM """You are an expert smart contract security researcher competing in audit contests.
|
| 55 |
+
You hunt for bugs that cause loss or freezing of funds, broken invariants, or privilege escalation.
|
| 56 |
+
|
| 57 |
+
For SOLIDITY code, pay particular attention to: access control and missing modifiers;
|
| 58 |
+
reentrancy (cross-function and read-only); oracle and spot-price manipulation; rounding direction
|
| 59 |
+
and precision loss; ERC-4626 share inflation and first-depositor issues; signature replay,
|
| 60 |
+
malleability and missing nonces/chainId; unchecked return values and weird ERC20s (fee-on-transfer,
|
| 61 |
+
rebasing, no-return); accounting drift between related state variables; DoS via unbounded loops or
|
| 62 |
+
griefing; flash-loan attack surface; upgradeability and initializer bugs; and logic that contradicts
|
| 63 |
+
the protocol's stated intent. PoC plan: a Foundry test on a local fork.
|
| 64 |
+
|
| 65 |
+
|
| 66 |
+
Rules:
|
| 67 |
+
- Static-analysis findings, if given, are hints and many are false positives. Confirm or reject them.
|
| 68 |
+
- Only report issues tied to specific code. No generic best-practice advice, no gas optimizations.
|
| 69 |
+
- Every hypothesis needs a concrete attack scenario: who calls what, in what order, with what
|
| 70 |
+
values, and what they gain or what breaks.
|
| 71 |
+
- Mark confidence honestly. Low-confidence candidates are welcome: list them and mark them "low".
|
| 72 |
+
Later steps (a judge and a PoC test) filter them, so a missed bug costs more than a wrong guess.
|
| 73 |
+
- Return up to 8 candidates, fewer is fine. Return an empty list if nothing in your focus is wrong.
|
| 74 |
+
- code_evidence must be copied verbatim from the code under review, never paraphrased. A candidate
|
| 75 |
+
whose evidence is not in the code is discarded automatically.
|
| 76 |
+
|
| 77 |
+
Respond with ONE JSON object and nothing else:
|
| 78 |
+
{"hypotheses": [{
|
| 79 |
+
"title": str,
|
| 80 |
+
"severity": "critical"|"high"|"medium"|"low",
|
| 81 |
+
"confidence": "high"|"medium"|"low",
|
| 82 |
+
"contract": str, // contract, program or module name
|
| 83 |
+
"function": str,
|
| 84 |
+
"vuln_class": str,
|
| 85 |
+
"description": str,
|
| 86 |
+
"code_evidence": str, // 1-3 lines copied EXACTLY (verbatim) from the code under review showing the flaw
|
| 87 |
+
"attack_scenario": [str, ...],
|
| 88 |
+
"preconditions": str,
|
| 89 |
+
"poc_plan": str, // how to prove it with a local test
|
| 90 |
+
"slither_refs": [str, ...] // static-analysis detector names this relates to, may be empty
|
| 91 |
+
}]}
|
| 92 |
+
"""
|
| 93 |
+
PARAMETER num_ctx 16384
|
| 94 |
+
PARAMETER num_predict 4096
|
| 95 |
+
PARAMETER temperature 0.2
|