AI Agent Hub
Back to plugins
🖥️

dsh-disk-audit

Client Updated 2026.08.25

Run the following command in DeepSeek Harness:

dsh plugin install zoahdev/dsh-disk-audit

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

Run dsh plugin install zoahdev/dsh-disk-audit in your DeepSeek Harness terminal to install; the full source is available at https://github.com/zoahdev/dsh-disk-audit .

About this plugin

DeepSeek Harness session logs are append-only. Leave them for a few weeks and they quietly balloon into hundreds of megabytes—or worse, hit the roughly 512 MB stringify cap before anyone notices. dsh-disk-audit exists for that exact moment: it tells you where the space went before it becomes a problem.

It scans your dsh data directory and produces a structured report: total size, per-directory breakdown, the largest directories, files exceeding a size threshold, and concrete cleanup suggestions. The whole process is read-only with zero runtime dependencies—it never deletes or moves a single file. Run it as a standalone CLI with semantic exit codes (0 clean, 1 warnings, 2 IO error), or let a harness agent call it directly so a natural-language question about data-directory usage returns a dsh-disk-audit/v1 report.

Built for developers who run dsh sessions long-term and have lost sight of how large their log directories have grown—especially on shared machines or in disk-constrained team environments. A quick audit in your routine is far cheaper than digging through several hundred megabytes of JSON after the fact.

Use Cases

  • Session logs silently balloon to hundreds of MB and you need to trace where space went
  • Routine audit of the dsh data directory to stay ahead of the stringify cap
  • Invoke from an agent in natural language to get a structured directory-usage report

Best For

  • Developers running dsh sessions long-term with continuously growing log directories
  • Engineering teams sharing machines with tight disk budgets
  • Users who want a lightweight disk-health check in their dsh workflow