← back to implementing-privileged-session-monitoring

SkillSpector · implementing-privileged-session-monitoring

independent scanner by NVIDIA · skill by mukul975 · how it works ↗

FAILmax severity: CRITICALrisk score: 100

Skill attempts to fill the context window with filler content, displacing legitimate instructions and safety constraints. This can degrade agent performance or ; Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.; Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.; +3 more

scanned 2026-07-07

Findings (12)

MEDIUMMemory Poisoningconfidence: 0.8

Skill attempts to fill the context window with filler content, displacing legitimate instructions and safety constraints. This can degrade agent performance or

SKILL.md

MEDIUMPrivilege Escalationconfidence: 0.8

Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.

SKILL.md

HIGHSupply Chainconfidence: 0.9

Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.

SKILL.md

HIGHSupply Chainconfidence: 0.9

Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.

SKILL.md

HIGHTool Misuseconfidence: 0.9

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glo

SKILL.md

HIGHTool Misuseconfidence: 0.85

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glo

SKILL.md

HIGHTool Misuseconfidence: 0.8

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glo

SKILL.md

HIGHTool Misuseconfidence: 0.7

Tool calls are chained to bypass individual safety checks or escalate capabilities beyond what any single tool call would allow.

SKILL.md

HIGHTool Misuseconfidence: 0.7

Tool calls are chained to bypass individual safety checks or escalate capabilities beyond what any single tool call would allow.

SKILL.md

HIGHYARA Matchconfidence: 0.75

YARA rule matched a known malware signature (reverse shell, backdoor, ransomware, C2 framework, or info stealer).

SKILL.md

CRITICALYARA Matchconfidence: 0.85

YARA rule matched a known malware signature (reverse shell, backdoor, ransomware, C2 framework, or info stealer).

SKILL.md

HIGHYARA Matchconfidence: 0.75

YARA rule matched a known malware signature (reverse shell, backdoor, ransomware, C2 framework, or info stealer).

SKILL.md

What the verdicts mean

SkillSpector reports on SkillMD's shared five-tier scale. See how SkillSpector works ↗.

PASS

Overall severity LOW (risk score in the safe range)

CAUTION

Overall severity MEDIUM

WARNING

Overall severity HIGH

FAILthis skill

Overall severity CRITICAL

INCONCLUSIVE

Scan could not complete