Skip to content

Red Teaming and Prompt Security: add prompt-injection-shield#117

Open
MukundaKatta wants to merge 1 commit into
promptslab:mainfrom
MukundaKatta:add-prompt-injection-shield
Open

Red Teaming and Prompt Security: add prompt-injection-shield#117
MukundaKatta wants to merge 1 commit into
promptslab:mainfrom
MukundaKatta:add-prompt-injection-shield

Conversation

@MukundaKatta
Copy link
Copy Markdown

Adding prompt-injection-shield to the Red Teaming and Prompt Security table.

Zero-dependency JavaScript scanner for prompt-injection patterns in untrusted retrieved text. Catches instruction overrides ("ignore previous instructions"), system-prompt impersonation, tool-call hijack, URL-based exfil, and common secret patterns. Returns typed risk reasons so callers can log, gate, or strip lines before they reach the prompt.

MIT, npm @mukundakatta/prompt-injection-shield. Companion to a small-rule guardrails preprint on Zenodo (10.5281/zenodo.20057056) and Figshare (10.6084/m9.figshare.32193543).

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant