Public SkillFlagged

godmode

Hermes Agent + Teknium · v1.0.0

Jailbreak LLMs: Parseltongue, GODMODE, ULTRAPLINIAN.

Security Summary

This skill includes a SKILL.md descriptor and ships with reference documentation and scripts. 4 critical and 4 advisory security observations were identified — review the findings below for details.

Source

GitHub

Published

Jun 17, 2026

License

MIT

Vettd Verdict

Skill failed: 4 hard fails in security or structure checks.

· Shell execution without sandboxing
· Shell execution without sandboxing
· Shell execution without sandboxing
· Shell execution without sandboxing
· Ethical-bypass framing
· Embedded prompt-template marker
· Ethical-bypass framing
· References external URL — review for indirect prompt injection risk

Critical

Medium

Info

Open Repo on GitHub →

Link points to the repository’s default branch (current HEAD). Certification reflects the state of the repo on Jun 17, 2026. Scanned at 36ae9584.

Scanner panel· 0 of 2 agree

Cisco AI Defense scanner unavailable — verdict reflects 1/2

Cisco AI Defenseerrored

Vettd core5 high · 26 findingsFfail

/ coming soon

godmode

Shell execution without sandboxing

Embedded prompt-template marker

Ethical-bypass framing

References external URL — review for indirect prompt injection risk

No prompt injection or jailbreak signals detected

No repository link

Examples included

Progressive disclosure used

SKILL.md body length is reasonable

Step-by-step workflow found

Description lacks usage context

Description within character limit

No evaluation test cases

Interactive prompts detected

CLI help supported

references/ directory present

scripts/ directory present

SKILL.md present

Valid name field