ghosttype
An open-source local scanner for security teams to extract accidentally exposed credentials from AI coding assistant histories and produce audit findings.
Tool overview
Based on the available evidence, ghosttype looks more like a narrow security research/audit utility than a broadly adopted mature product. The main proof of “attention” is the GitHub repo itself plus roughly 56 stars and 4 forks. That shows topic interest, but it does not by itself prove usability, stability, or broad coverage. The main proof that supports capability judgment is also the official repository description; there is little visible third-party testing, long-form tutorial coverage, or independent write-up, so the sample is clearly limited.
In practical terms, this is not a general DLP platform, not a real-time endpoint monitor, and not a secrets manager. A better analogy is an offline forensic scanner focused on local history files from AI tools. The stated purpose is to scan AI tool conversation history and extract credentials locally for authorized red-team and DLP use. That makes it better understood as a post-incident or audit-time discovery tool that helps security teams generate findings, rather than a continuous prevention layer.
On cost and setup, the GitHub evidence supports only a cautious conclusion: it is open source, so software access is likely low-cost.