AI Agent Hub
Back to plugins
⚙️

dsh-hooks-pack

Workflow Updated 2026.08.25

Run the following command in DeepSeek Harness:

dsh plugin install chenzhi-clude/dsh-hooks-pack

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

Install in DeepSeek Harness by running dsh plugin install chenzhi-clude/dsh-hooks-pack; source code is available at https://github.com/chenzhi-clude/dsh-hooks-pack

About this plugin

If you have already built hooks in Claude Code or Codex and want them to work the same way inside DeepSeek Harness, dsh-hooks-pack is that bridge. It auto-discovers existing hook configs at standard paths, mounts both dialects onto the DSH host process in a single step, and—when nothing is found—registers nothing at all, so a fresh install never silently changes agent behavior.

Beyond carrying over your scripts, the pack ships a ready-to-use starter set: guard-dangerous intercepts shell calls that attempt rm -rf, force pushes, fork bombs, or disk formatting before they execute; audit-log appends every tool invocation to a JSONL file for after-the-fact review. Together they give you a working safety net in under a minute.

This plugin suits two profiles: developers who have accumulated hooks in the Claude Code or Codex ecosystem and want a friction-free migration to DSH, and security-minded users who want dangerous-command guarding and call auditing up and running without hand-writing configuration from scratch.

Use Cases

  • Migrate existing Claude Code or Codex hooks into DSH without rewriting any code
  • Turn on dangerous-command guarding (rm -rf, fork bombs, disk formatting, etc.) in one step
  • Append every tool invocation to a JSONL audit trail for after-the-fact review in DSH workflows

Best For

  • Developers with existing Claude Code or Codex hook setups looking to move to DSH
  • Security-minded users who want instant dangerous-command guarding without hand-writing configs
  • Team leads who need audit trails for every DSH tool-call invocation