dsh-plugin

by cromus-ai

GitHub ↗ npm ↗

0 技能包github 检测到 manifest package.json#dsh收录于 08-16

DeepSeek Harness包：使用Cromus评分、验证和成本估算AI技能

DeepSeek Harness bundle: score, validate, and cost-estimate AI skills with Cromus.

安装

$ dsh plugin --profile web add github:cromus-ai/dsh-plugin

GitHub 源码安装：首次需按提示配置 allowBuilds 构建授权后重试

安装与环境配置指引、插件开发教程见 DSH 中文社区文档 ↗

安装即在你的机器上以你的权限运行第三方代码——它可读写文件、使用凭据、访问网络，DSH 的工具审批不会为插件代码加沙箱。「检测到 manifest」仅代表发现 dsh.bundle / dsh.plugin 清单，不构成兼容性或安全审查；安装前请审阅源码，不熟悉的插件先在不含密钥的环境试用。

Cromus plugin for DeepSeek Harness — score, validate, and cost-estimate AI skills directly from the dsh agent.

Install

dsh plugins add @cromus-ai/dsh-plugin

Then set your API key (get one at cromus.ai/connect):

export CROMUS_API_KEY=your_key_here

Tools

Once installed, all 11 Cromus tools appear in the model's tool registry under the mcp__cromus__ namespace:

Model-facing name	What it does
`mcp__cromus__score_skill`	Returns a 0–1000 Croms score with a four-dimension breakdown for any skill file
`mcp__cromus__simulate_cost`	Estimates per-run and monthly cost across five routing modes for a given workflow
`mcp__cromus__check_mcp_drift`	Compares a skill's declared MCP dependencies against a live server and flags changes
`mcp__cromus__check_model_drift`	Detects retired, renamed, or superseded models declared in a SKILL.md
`mcp__cromus__validate_skill`	Checks a SKILL.md file against the open specification and lists any violations
`mcp__cromus__validate_mcp_deps`	Checks a skill's `mcp_dependencies` block for completeness and valid configuration
`mcp__cromus__validate_ethos`	Checks an ETHOS.md governance file against the open specification
`mcp__cromus__validate_memory`	Checks a MEMORY.md handoff file against the open specification
`mcp__cromus__validate_agents`	Validates an AGENTS.md file and returns a readiness score
`mcp__cromus__reality_check_summary`	Returns the most recent Reality Check record for a skill, read-only
`mcp__cromus__check_usage_waste`	Analyzes AI API usage exports for waste patterns: cache utilization, model sprawl, idle burn

Manual config

If you prefer to wire it in yourself, add this to your cordis.patch.yml:

- insert:
    - id: cromus-mcp
      name: '@deepseek-ai/dsh-mcp-client'
      config:
        serverName: cromus
        transport: streamable-http
        url: https://mcp.cromus.ai/mcp
        headers:
          Authorization: !!js '`Bearer ${process.env.CROMUS_API_KEY}`'

License

MIT

原始 README： https://github.com/cromus-ai/dsh-plugin/blob/main/README.md ↗

同类插件

查看全部 →

技能包

vibe-skills

vibe 编码技能合集（含快速开始、CI 证明、发布说明与文档）

查看详情

2943github+08-23

技能包

Aegis

面向编码 Agent 的软件工程方法包，提供基线优先规划、系统化调试、提示词卫生、完成前验证，以及修复/退役双轨跟踪技能。

查看详情

1020github+08-16

技能包

J-Space-Cognition-Suite-V3.6

A model-agnostic inference-time control suite for deep reasoning, long-horizon work, tool use, verification, and recovery. 面向深度推理、长程任务、工具调用、验证与恢复的模型不可知推理时控制套件。

查看详情

873github+08-17

技能包