marcostolosa/taintzero ? reverse-engineered prompt
Reverse engineered prompt
Build me a command line tool called TaintZero that scans a local folder or a GitHub repo for security issues using Semgrep, then optionally sends the findings to an LLM for triage.
I want it to detect languages, frameworks, and manifests when possible, work on Linux and Windows, and produce a clean report with the raw Semgrep results plus a separate triage file when LLM review is enabled. The LLM part should read a YAML config, support local or cloud providers, and only send code out if I explicitly allow cloud triage. Please redact secrets before anything leaves the machine.
Make the CLI feel polished, with tables, progress, and sensible status messages. Include commands to scan a repo, triage an existing report, list providers, and sync rules from Semgrep Registry or local rules. Support offline mode with cached rules, remote Git clones that are shallow and non interactive, and sensible defaults like auto rules selection. If you need to, look up current docs online so the provider setup and Semgrep usage are correct.
Are you gonna build this?
make sure you review the code using coderabbit