jmanico/grim-prompter
Grim Prompter guides coding agents through threat models backed by file-and-line evidence.
The repository provides 36 architecture workflows and 29 method and analysis prompts, with quick reviews and a six-stage deep workflow. Dedicated privacy reviews examine data collection, sharing, retention and potential harms to people. Its runbook requires read-only inspection; executing tests, installs or network requests needs separate authorization. Reports include evidence gaps, proposed mitigations and verification plans, and saved JSON models support later reviews of code changes.
GitHub · LLM repos' note
The repository provides 36 architecture workflows and 29 method and analysis prompts, with quick reviews and a six-stage deep workflow. Dedicated privacy reviews examine data collection, sharing, retention and potential harms to people. Its runbook requires read-only inspection; executing tests, installs or network requests needs separate authorization. Reports include evidence gaps, proposed mitigations and verification plans, and saved JSON models support later reviews of code changes.
GitHub · LLM repos' note
score 4