protectai/llm-guard

B-Tier scored 2026-07-07 metrics 2026-07-07

A security toolkit for LLM interactions: scanners for prompt injection, PII, toxicity, and secrets, applied to inputs and outputs. Agentiquette scores it 68/100 (B-tier), methodology v1.0, scored 2026-07-07.

68/100 B-Tier
scored 2026-07-07 · methodology v1.0 · how we score
Reliability6825%
Skill Leverage7020%
Documentation7615%
Maintenance4010%
Safety / Governance8610%
Evaluation Readiness6610%
Composability745%
Adoption (capped)585%

Key facts

CategoryGovernance / guardrails
LicenseMIT
MaintainerProtect AI
Compatible agentsgeneric
File patterns
Triggeringmanual-reference
GitHub metrics3,152 stars · 416 forks · 37 open issues (fetched 2026-07-07)
Last commit2025-12-15
Created2023-07-27

Editorial analysis

Security-team DNA rather than ML-team DNA, visible in the scanner taxonomy: injection, secrets, and leakage treated as the attack surface they are. The maintenance slowdown after the corporate acquisition is the material fact; injection defense is an arms race, and stale scanners decay silently. Watch status.

Risks

Seven-month commit gap at fetch time following Protect AI's acquisition; scanner freshness against new injection techniques is the concern.

Evidence notes: Safety: scanner coverage is broad and well-categorized. Maintenance: 7-month gap at fetch; flagged for staleness watch. Composability: plain scanner API, drops into any pipeline.

Install and first run

install · not yet run by us
pip install llm-guard

Alternatives in this category

Compare with guardrails-ai

Embed this badge

Maintainers of scored entries can embed their live badge; it updates when the score does and links back to this page.

Agentiquette: B tier, 68/100

README.md
[![Agentiquette](https://agentiquette.com/badge/llm-guard.svg)](https://agentiquette.com/index/repos/llm-guard)

Frequently asked questions

Is protectai/llm-guard good?

protectai/llm-guard scores 68/100 (B-tier) on Agentiquette's methodology v1.0, scored 2026-07-07. Strengths and weaknesses are in the editorial analysis; the score is explained dimension by dimension in the scorecard.

What is protectai/llm-guard for?

A security toolkit for LLM interactions: scanners for prompt injection, PII, toxicity, and secrets, applied to inputs and outputs.

Is protectai/llm-guard maintained?

Last commit 2025-12-15, 37 open issues at fetch time (2026-07-07). Maintenance is scored at 40 of 100.

Maintainer of protectai/llm-guard? Claim this listing and suggest corrections.