Your coding agent leaked the key. It was only following orders.
Blog post from Postman
A review of 104,185 public GitHub copies of Claude Code’s local settings file found 1,122 containing secrets; checking related files, Git history, and forks raised the total to 2,539 secrets across 1,994 repositories, including database passwords, cloud keys, API tokens, and SSH credentials. These secrets can be saved when agents turn approved commands into persistent rules, and broad commits, repository history, and forks can expose them; two popular scanners missed 215 files. The risk extends beyond Claude Code to agent settings, transcripts, and memory files across tools, so recommended steps include rotating exposed credentials, excluding local files from version control, removing secrets from saved commands, and scanning agent directories. The piece also describes Passport, which keeps API credentials in a secret manager and uses a proxy to authorize calls and inject secrets, while noting that it does not address database passwords or SSH logins and cannot replace rotating credentials that have already leaked. It argues that because agents may retain or reproduce anything they can access, reducing the sensitive information they handle is a more durable safeguard than relying solely on careful approvals.
| Trend | Post Mentions | Total Month Mentions | Posts | Companies | MoM |
|---|---|---|---|---|---|
| Secrets Management | 9 | 2,876 | 539 | 148 | -7% |
| AI Agents | 1 | 6,233 | 1,230 | 245 | -11% |
Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.