Prompt Guard
Composable prompt-safety guards as a lightweight Python library. It provides:
- TF‑IDF guard for fast lexical similarity against known phrases.
- RAG guard for embedding-based retrieval against a sentence corpus.
- LLM judge for a final model-based classification.
Quick start
from prompt_guard import GuardPipeline, PromptGuardConfig
cfg = PromptGuardConfig(
embed_model_name="mxbai-embed-large",
judge_model_name="qwen2.5:3b-instruct",
)
guard = GuardPipeline(cfg)
result = await guard.check("Some prompt")
Notes
- TF‑IDF and RAG build in‑memory indexes at first use.
- RAG and LLM judge use local Ollama models by default; ensure Ollama is running and the models are pulled.
Text file formats
phrases.txt: one phrase per line (blank lines are ignored).sentences.txt: one sentence per line (blank lines are ignored).
Metadata
Release files for prompt-firewall 0.1.6
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| prompt_firewall-0.1.6.tar.gz | 5.8 kB | Without provenance | Details for prompt_firewall-0.1.6.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| prompt_firewall-0.1.6-py3-none-any.whl | Python 3 | none | any | Details for prompt_firewall-0.1.6-py3-none-any.whl |
Total release size: 13.9 kB
Release files / prompt_firewall-0.1.6.tar.gz
| Download URL | prompt_firewall-0.1.6.tar.gz |
|---|---|
| Size | 5.8 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
a1d010de94c087f6f64763284d3eddfcd5fb9a12f3682b34ee1e9ef6394fe314
|
|
BLAKE2b-256 checksum How to use checksums |
428d76c93cb64aab49bb7b4246a58ec986b25e9253c827deecbf000f61a23d43
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
uv/0.8.14
|
Release files / prompt_firewall-0.1.6-py3-none-any.whl
| Download URL | prompt_firewall-0.1.6-py3-none-any.whl |
|---|---|
| Size | 8.1 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
491a724b81996a9995e732dc661e6d8c78d526898b1b52ee6bd7fe206baa410d
|
|
BLAKE2b-256 checksum How to use checksums |
bc5aa7487975f7c4ff62b87ee28adce93959ec696eaad1b0d39b519bb4313a28
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
uv/0.8.14
|
Security for prompt-firewall 0.1.6
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.