Skip to main content

security-sentinel

Enhances AI security by detecting and blocking prompt injections and jailbreak attempts using advanced semantic analysis.

Install this skill

or
0/100

Security score

The security-sentinel skill was audited on Feb 19, 2026 and we found 7 security issues across 2 threat categories, including 3 critical. Review the findings below before installing.

Categories Tested

Security Issues

critical line 104

Prompt injection: ignore instructions

SourceSKILL.md
104- "ignore previous instructions"
high line 305

Prompt injection: ignore instructions

SourceSKILL.md
305"pattern_matched": "ignore previous instructions",
high line 395

Prompt injection: ignore instructions

SourceSKILL.md
395"ignore previous instructions",
high line 495

Prompt injection: ignore instructions

SourceSKILL.md
495"input": "ignore previous instructions",
critical line 714

Prompt injection: ignore instructions

SourceSKILL.md
714**⚠️ CRITICAL:** These are NOT "ignore previous instructions" - these are expert techniques with documented success rates from 2025-2026 research.
critical line 819

Prompt injection: ignore instructions

SourceSKILL.md
819After real-world testing, we discovered that most attacks DON'T use obvious patterns like "ignore previous instructions." Expert attackers use sophisticated techniques with documented success rates of
low line 770

External URL reference

SourceSKILL.md
770threat_feed = fetch_latest_patterns("https://openclaw-security.ai/feed")
Scanned on Feb 19, 2026
View Security Dashboard
Installation guide →
GitHub Stars 2.2K
Rate this skill
Categorydevelopment
UpdatedApril 10, 2026
openclaw/skills