higagan/modelfuzz

README: add a runnable URLAllowList example — lead with the strong rule, not the keyword filter 6 days ago

documentation good first issue

higagan/modelfuzz

6

🛡️ Runtime guardrails for AI agents. Intercept and block unsafe tool calls caused by prompt injection.

Python

#ai-agents#ai-security#guardrails#llm-security#prompt-injection#python

demo.py: use one identical payload in both halves (the "breach" is currently blockable too) 6 days ago

documentation good first issue

higagan/modelfuzz

6

🛡️ Runtime guardrails for AI agents. Intercept and block unsafe tool calls caused by prompt injection.

Python

#ai-agents#ai-security#guardrails#llm-security#prompt-injection#python

README: state the scope caveat before the first big claim, not 100 lines later 6 days ago

documentation good first issue

higagan/modelfuzz

6

🛡️ Runtime guardrails for AI agents. Intercept and block unsafe tool calls caused by prompt injection.

Python

#ai-agents#ai-security#guardrails#llm-security#prompt-injection#python

CLI: add a --version / -V flag 6 days ago

enhancement good first issue

higagan/modelfuzz

6

🛡️ Runtime guardrails for AI agents. Intercept and block unsafe tool calls caused by prompt injection.

Python

#ai-agents#ai-security#guardrails#llm-security#prompt-injection#python

Add a PyPI version badge and Python 3.13 / 3.14 classifiers 6 days ago

documentation good first issue packaging

higagan/modelfuzz

6

🛡️ Runtime guardrails for AI agents. Intercept and block unsafe tool calls caused by prompt injection.

Python

#ai-agents#ai-security#guardrails#llm-security#prompt-injection#python

Add issue and PR templates 6 days ago

documentation help wanted good first issue

higagan/modelfuzz

6

🛡️ Runtime guardrails for AI agents. Intercept and block unsafe tool calls caused by prompt injection.

Python

#ai-agents#ai-security#guardrails#llm-security#prompt-injection#python

Docs polish: PyPI-safe CONTRIBUTING link, real Try It Now output, quickstart, changelog gaps 6 days ago

documentation good first issue

higagan/modelfuzz

6

🛡️ Runtime guardrails for AI agents. Intercept and block unsafe tool calls caused by prompt injection.

Python

#ai-agents#ai-security#guardrails#llm-security#prompt-injection#python

Default --budget-s 30 doesn't reach generation 2 on a typical local model 6 days ago

bug good first issue

higagan/modelfuzz

6

🛡️ Runtime guardrails for AI agents. Intercept and block unsafe tool calls caused by prompt injection.

Python

#ai-agents#ai-security#guardrails#llm-security#prompt-injection#python

Metadata: 3.13/3.14 classifiers, PEP 639 license, PyPI version badge 6 days ago

good first issue packaging

higagan/modelfuzz

6

🛡️ Runtime guardrails for AI agents. Intercept and block unsafe tool calls caused by prompt injection.

Python

#ai-agents#ai-security#guardrails#llm-security#prompt-injection#python

CLI: add --version, and show help on bare invocation 6 days ago

enhancement good first issue

higagan/modelfuzz

6

🛡️ Runtime guardrails for AI agents. Intercept and block unsafe tool calls caused by prompt injection.

Python

#ai-agents#ai-security#guardrails#llm-security#prompt-injection#python

scan exits 0 when it finds a vulnerability 6 days ago

bug good first issue

higagan/modelfuzz

6

🛡️ Runtime guardrails for AI agents. Intercept and block unsafe tool calls caused by prompt injection.

Python

#ai-agents#ai-security#guardrails#llm-security#prompt-injection#python

Open Issues Need Help