dsh-plugin
Run the following command in DeepSeek Harness:
dsh plugin install cromus-ai/dsh-plugin
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install cromus-ai/dsh-plugin inside DeepSeek Harness to add the Cromus governance plugin; see https://github.com/cromus-ai/dsh-plugin for full integration details.
About this plugin
As AI skill files pile up across a team, model declarations in SKILL.md quietly go stale, MCP dependencies drift away from live services, and API bills silently harbour waste nobody notices — problems that are hard to spot during everyday dsh agent runs. dsh-plugin injects eleven Cromus governance tools into the dsh agent tool registry, covering a four-dimension skill score on a 0-1000 scale, line-by-line spec validation for SKILL.md, ETHOS.md, MEMORY.md, and AGENTS.md, MCP and model drift detection, per-run and monthly cost estimation across five routing modes, and a dedicated waste-pattern audit flagging low cache utilisation, model sprawl, and idle burn in API usage exports.
It is aimed at teams that manage AI skill workflows with DeepSeek Harness, especially developers who maintain skill files across Claude Code, Cursor, Cline, and Codex and need a consistent governance baseline plus ongoing control over inference costs.
Use Cases
- Validate and score skill files directly from the dsh agent tool registry
- Detect whether models declared in SKILL.md have been retired, renamed, or superseded
- Audit API usage exports for wasteful cache utilization, model sprawl, and idle burn
Best For
- Developers managing AI workflows with DeepSeek Harness
- Teams maintaining skill files across Claude Code, Cursor, Cline, and Codex
- Platform operators focused on controlling and optimizing AI inference spend
Related Plugins
A unified suite combining hot runtime injection, task-aware thinking-mode routing, and a graded session protocol with red-team gates to sustain model diligence across long-horizon inference.
ModLens is a vision plugin for DeepSeek Harness that gives text-only models sight by reading images pasted directly into chat, with zero-config setup and multiple vision engines.
On-demand vision for text-only DeepSeek Harness agents: built-in free keyless vision chain and 14 vision tools, routing image turns as tool calls to vision models with pixel fidelity, no Python needed, one-command install.
Give text-only models in DeepSeek Harness eyes, enabling image Q&A, long-screenshot OCR, UI restoration, and GUI visual tasks.