General skillSecurity and privacy
prompt-injection-review
Traces prompt-injection paths from untrusted inputs to dangerous actions in LLM applications before release.
Traces prompt-injection paths from untrusted inputs to dangerous actions in LLM applications before release.
prompt-injection-review maps trust boundaries in LLM applications that use RAG, external content, tools, agents, or model-rendered output, tracing untrusted sources through the model to outbound actions, execution, data access, or UI sinks. It recommends least privilege, human approval, egress controls, structured validation, and authorization enforced in code. It explicitly treats prompt injection as having no complete fix: the goal is to contain what a successful injection can reach, not to rely on filters for false certainty.
Resource types
Use cases
Platforms
Capabilities
Audit suspension risks in a Claude Code working environment.
Check whether a specific CVE actually applies to your environment.
Public GitHub facts last synced Jul 10, 2026.
Store and inject development secrets through 1Password from Claude Code.
Audit a deployed public app for production gaps missed during coding.