skills sourced live from GitHub · all skills listed are free · each skill is provided under its author's license.
home / Security / Prompt Injection Defense Skill by @BagelHole
The skill is the security/ai/prompt-injection-defense folder inside that repository. Copy it into the folder your agent loads skills from; check your agent's documentation for where that is.
Layered defences against direct and indirect prompt injection: input sanitisation, canary tokens, tool permissions, output validation and isolation.
Maps attack surfaces and gives code and configuration for defence in depth, red-team cases and monitoring queries.
Written and maintained by @BagelHole. It is published under the MIT license, which allows reuse with attribution. See the author profile. SkillDirectory does not claim authorship: this page adds our own review and links back to the original, and the skill's text belongs to its author.
(Excerpt. Full file: https://github.com/BagelHole/DevOps-Security-Agent-Skills/blob/main/security/ai/prompt-injection-defense/SKILL.md)
Excerpt only, quoted from the author's repository. Open the repository for the complete file.