SkillAgentSearch skills...

dsh-rea

DeepSeek Harness plugin bundle: REA (Reverse Engineer Anything) MCP tools behind a single Plugins-page switch. Ships no engine — only the connection declaration.

Install / Use

claude mcp add PandaBoby -- npx -y github:PandaBoby/dsh-rea

If the server publishes to npm under a different name, use that package instead — check the repo README.

About this skill
🔌

MCP Server

Model Context Protocol server

Quality Score

78/100

Supported Platforms

Claude Code
Claude Desktop

Tags

Our assessment of dsh-rea

dsh-rea scores 78/100 on our quality scale, 3708th of 4,705 Development & Engineering skills we index.

Its MCP Server is 4.7 KB long, well organised into 9 sections with 4 code examples: a solid amount of guidance for an agent.

It has 3 GitHub stars, so there is little community track record yet; judge it on its content.

Substance
26/30
Structure
20/20
Description
15/15
Adoption
3/20
Freshness
15/15

Maintenance, license and trust

  • The repository was last updated 28 days ago, so dsh-rea is actively maintained.
  • It is released under the MIT license, a permissive license that allows use, modification and commercial use with attribution.
  • Its trust signals score 92/100, with 1 caution from licensing, adoption, age or documentation. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.

Safety scan

No issues found

Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.

AI review by kimi-k2.7-code on 2026-11-16. Automated pattern scan on 2026-11-16. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.

dsh-rea compared with similar skills

All 4 of these similar skills score higher than dsh-rea; compare them before choosing.

SkillScoreStarsUpdatedFormat
dsh-rea (this skill)by PandaBoby78328d agoMCP Server
taste-skillby Leonxlnx100117.0k8d agoCLAUDE.md
Agent-Reachby Panniantong100112.4k9d agoCLAUDE.md
rufloby ruvnet10081.9ktodayCLAUDE.md
headroomby headroomlabs-ai10079.9ktodayCLAUDE.md

Frequently asked questions

How do I install dsh-rea?
Run claude mcp add PandaBoby -- npx -y github:PandaBoby/dsh-rea. The install tabs above show the steps for each supported agent.
Which AI agents does dsh-rea work with?
It is written for Claude Code and Claude Desktop, as a MCP Server file. Other agents that read the same format can often use it too.
Is dsh-rea safe to use?
Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It is MIT-licensed and scores 92/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
Is dsh-rea still maintained?
The repository was last updated 28 days ago, so dsh-rea is actively maintained.

dsh-rea

A DeepSeek Harness plugin bundle that mounts the REA (Reverse Engineer Anything) MCP server as the rea tool namespace — and gives it an ordinary on/off switch in the Plugins page.

This package ships no analysis engine, no tool implementation, and no model code. It is a declaration: one MCP bridge row, plus the display metadata that makes it a card in the plugin list. Every actual REA tool is discovered from the rea-agents package at run time.

Why a bundle instead of a YAML snippet

Dropping the same six lines into ~/.dsh/cordis.patch.yml works, but that row is invisible to the Plugins page — the page lists bundles (packages that carry a dsh.bundle.patch), not free-standing rows. Wrapping the row in a bundle is what turns it into a switch you can click, or click back off after the investigation.

The switch matters because the rea namespace is expensive while it is on: about 200 tool schemas travel in every request.

| State | Tools sent | Tool-definition context | | --- | --- | --- | | dsh-rea off | ~40 built-in | ~2K–8K tokens | | dsh-rea on | ~40 + ~207 | ~200K tokens |

Measured against rea-agents@6.5.0 on a 1M-token window: the schemas serialize to ~790 KB, which the harness prices at ~207K tokens, roughly 20% of the window. The figure moves with the REA release; it is never zero while the bundle is on.

Install

From GitHub:

dsh plugin add github:PandaBoby/dsh-rea

Or from a local checkout, which is the form to use while developing this package:

dsh plugin add /absolute/path/to/dsh-rea

Installation also selects the bundle, so rea tools appear right away; switch it off in the Plugins page when you are done. Grab the matching skill from rea-agents (or from the REA repository) if you want the model to have REA's own procedure guidance; this package deliberately does not duplicate it.

The REA server itself is not downloaded at install time. The bridge runs npx -y rea-agents@latest mcp, so npm fetches the newest release on the first tool sync after the switch is turned on (and reuses its cache afterwards).

Turn it off and on

Plugins page → Installed plugins → dsh-rea — flip the switch.

  • On — the MCP bridge starts, lists REA's tools, and registers them under mcp__rea__*.
  • Off — the row is withdrawn, the child process is disposed, and the tool schemas stop being sent. Later requests are ~200K tokens lighter.

The switch is a profile-level setting: it applies to every session of the profile, including sessions that are already open. Turning it off while a REA call is running cancels that call.

No restart is needed when HMR is enabled (the default in Desktop and Web profiles). Without HMR, the change applies on the next start.

Configuration

Defaults live in cordis.patch.yml. Override the rea-mcp row from your profile patch when you need to pin a REA version, raise the call timeout for slow decompilation, or make startup failures loud:

- id: rea-mcp
  config:
    serverName: rea
    transport: stdio
    command: npx
    args: ["-y", "rea-agents@6.5.0", "mcp"]
    toolCallTimeoutMs: 600000
    failOnStartupError: false

A patch replaces a row's whole config rather than merging into it, so restate every key you want to keep.

| Key | Default | Meaning | | --- | --- | --- | | serverName | rea | Namespace of the tool names (mcp__rea__<tool>); max 32 chars | | command / args | npx + -y rea-agents@latest mcp | The stdio server to launch | | toolCallTimeoutMs | 60000 | Per-call budget; raise it for long analyses | | failOnStartupError | false | true fails the plugin instead of silently exposing no tools | | reconnect | built-in policy | Retry/backoff when the child process drops |

Verify what it costs

The context meter above the composer breaks usage into system prompt, tool definitions, and messages. Toggle the bundle and compare the tool definitions row — that delta is what this plugin costs. Alternatively, read the latest request/header from a session log under ~/.dsh/sessions/ and count the mcp__rea__* entries.

Uninstall

Plugins page → dsh-rea → Remove. Removing the bundle unregisters the row and lets you delete the package; the npx cache is outside the profile and stays until npm or npx clears it.

Layout

package.json        name, description, icon, dsh.bundle.patch
cordis.patch.yml    the one insert row: mcp-client connected to `rea mcp`
icon.svg            card artwork
locale/en.json      card title + description
locale/zh.json      same, Chinese

Licence

MIT. REA itself is a separate project with its own licence; this package only declares how to reach it.

Related Skills

View on GitHub
GitHub Stars3
CategoryDevelopment
Updated28d ago
Forks0

Trust signals

92/100

From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.

1 low