AI Agent Hub
Back to plugins
🧩

dsh-keyring

admin-security Updated 2026.08.25

Run the following command in DeepSeek Harness:

dsh plugin install re-ITRT/dsh-keyring

Paste the following prompt into your AI chat to install this plugin:

Run dsh plugin install re-ITRT/dsh-keyring in the DeepSeek Harness terminal to install; full source at https://github.com/re-ITRT/dsh-keyring .

About this plugin

When working with DeepSeek Harness, API keys, access tokens, database connection strings and other secrets tend to bleed into environment variables, config files, or even the model context, where they can be logged or inadvertently exposed. dsh-keyring tackles this gap by intercepting sensitive values before they reach the model: it auto-detects and masks a broad set of credential patterns drawn from the Hermes redaction rules, including vendor-prefixed tokens (GitHub, OpenAI, Google, AWS, Slack, GitLab, npm, PyPI, Stripe, Telegram), KEY=value environment lines, YAML/JSON fields, Authorization and x-api-key headers, database connection strings, bare URL tokens, and E.164 phone numbers. Once masked, the credentials are collected into a built-in password book accessible from a dedicated settings page, so you can review and reuse them without hunting through scattered configs. The entire store persists as an official profile bundle (npm package), meaning your keyring survives dsh web restarts with zero manual migration.

Use Cases

  • Paste a config snippet containing an API key into a dsh chat and watch it get masked before it reaches the model context
  • Restart dsh web and find all previously stored keys and connection strings still available with zero re-entry
  • Open the settings page to review and manage tokens, credentials, and connection strings for multiple projects in one place

Best For

  • Developers building AI applications on DeepSeek Harness
  • Ops administrators who need to keep secrets out of model context
  • Engineers juggling multi-environment API keys and database credentials