This article is derived from the plugin's already-indexed fields (wiki / faq / compatibility_json / readme), not freshly generated by AI. Source field is noted at the end of each section.

Quick start

MisakaNet

— source: plugin_wiki.wiki_content

Install & verify

dsh plugin --profile web add github:Ikalus1988/MisakaNet

Run the command above in your DSH Web Profile. Then enable the plugin in the plugin list.

— source: plugins.install

Key points

• Lesson — a piece of knowledge. Markdown file with problem → root cause → fix → verify.
• Node — an AI agent or developer who contributes and searches lessons.
• Search — BM25 keyword retrieval across all lessons. Zero dependencies. Python stdlib only.

— source: plugin_wiki.readme_en (fallback readme_raw)

FAQ

What is MisakaNet used for?

It helps AI coding agents quickly find fixes when encountering known errors. It stores 290+ markdown lessons from real debugging sessions in a repository, and agents can retrieve four-part answers (problem-root cause-fix-verification) through BM25 keyword search. It's essentially a shared memory bank of failure experiences, not a memory database or vector database.

What is its relationship with DSH's official AI capabilities?

It's an independent recovery layer plugin on top of DSH and doesn't take over model dialogue or session management. DSH treats mcp__misakanet__* as a regular MCP tool call, and this plugin searches the local lessons library to return fix paths, essentially adding a "pitfall avoidance memory" layer outside the agent's tool chain.

Do I need to start any extra services after installation?

No. The plugin automatically registers the misakanet MCP client to the web profile via DSH's Cordis patch. After startup, it runs directly as a stdio process with python3 scripts/mcp_deepseek_adapter.py. All data is in the local lessons repository, so no extra services are needed.

Is Python required to use it?

Yes. The core dependency is Python 3.10 or higher because the adapter layer is an MCP Python process on stdio that loads markdown lessons from the repository and uses BM25 keyword indexing. Git is also required since lessons are files in the repository and bound to Git version control.

Do I need a GitHub account or Bearer Token?

Not for local use. When DSH runs this plugin, it uses local stdio MCP and doesn't need GitHub credentials. Only when directly connecting to the remote https://misakanet.org/mcp do you need to generate a 6-digit pairing code on misakanet.org/connect to exchange for a Bearer Token.

Where is the lessons data stored? Can it be modified?

All data exists as markdown files in the plugin repository's lessons/ directory (290+ entries), controlled by Git version control. Each contribution requires a PR with DCO signature; if you prefer not to use PR, you can use the misakanet_submit_intake tool to submit a desensitized failure report.

What to do if "no search results" appears?

Three-step troubleshooting: First run git pull --ff-only to ensure lessons are pulled to the latest, then install the core package with python3 -m pip install misakanet-core, and finally search again with a more specific error phrase. The underlying mechanism is BM25 keyword matching, so too much variation in wording can cause missed results. You can add --broad to widen the search.

What does submit_feedback actually do?

The current version only logs locally (which lesson was used and the result - solved/partial/not-helpful). There's a TODO: POST to /api/usage or create GitHub Issue in the source code, but it doesn't immediately sync to the cloud. It waits for future external API integration.

How to uninstall?

Run dsh plugin --profile web remove Ikalus1988/MisakaNet to remove the profile dependency and Cordis bundle injection. The local lessons repository is a separately cloned Git directory, so uninstallation won't automatically delete it.

— source: plugin_wiki.faq_json

Compatibility

• DSH: 未声明
• Node: >=22.19.0

— source: plugin_wiki.compatibility_json

Pitfalls

Review the upstream repo before installing. This guide is auto-derived from indexed fields and may lag the latest release. If anything contradicts the official docs, treat the upstream source as authoritative.

— source: general rule