編集者注

插件全景

Know what exists before you build what is missing.

これは DeepSeek Harness(DSH)プラグインです。当サイトは GitHub README、インストール情報、メンテナンス状況、公開セキュリティシグナルをまとめています。

上流で中国語 README が提供されていないため、リポジトリのオリジナルコンテンツを表示しています。

DSH Landscape

简体中文 · Roadmap · Contributing

CI Node.js >= 22 License: MIT

Know what exists before you build what is missing.

DSH Landscape answers one practical question: given a capability need, what already exists in the DeepSeek Harness ecosystem, what is still missing, and should you use, extend, investigate, or build?

It combines plugin discovery, evidence-based maturity, coverage-aware verdicts, and fresh verification. A repository name or README claim alone never counts as proof that a capability is solved.

What you get

  • A native DSH tool: dsh_landscape.
  • Need-first analyze, find, and brief CLI workflows.
  • A portable Agent Skill for Codex, DSH, and other compatible hosts.
  • Structured evidence with provenance, maturity, freshness, uncertainty, and next-action guidance.

Quick start: install as a DSH plugin

Install the verified, reproducible revision into your DSH profile:

dsh plugin --profile web add github:cyanseek/dsh-landscape#1ef0e1ebbb5e84e8d2feed31ae71d1b97322f6f9
dsh --profile web --dump-config

The bundle registers dsh_landscape. Ask DSH a need such as:

Check whether DSH already has a native Linear integration. Show the evidence and tell me whether to use, extend, or build.

No separate Landscape model key is required inside DSH. The tool retrieves and verifies evidence; the host model makes the final semantic decision.

Runtime options

ArgumentDefaultMeaning
needrequiredCapability or integration described in natural language
limit10Maximum matching projects, from 1 to 20
freshtrueLive-check negative or uncertain results through GitHub search

There is no required plugin configuration. Anonymous GitHub search works out of the box; an existing GITHUB_TOKEN can increase API limits. Its value is never included in tool output.

Uninstall

dsh plugin --profile web remove dsh-landscape

DSH unregisters the tool when the bundle is removed. Replace web with the profile you actually use.

Use as an Agent Skill

Use once without permanent installation:

npx -y skills use cyanseek/dsh-landscape --skill dsh-landscape --agent codex

Or install the portable Skill:

npx -y skills add cyanseek/dsh-landscape --skill dsh-landscape -g -a codex -y

The generic prompt-only form works with other capable hosts by omitting --agent codex.

Run the CLI

The GitHub-source commands work without an npm release:

npx -y github:cyanseek/dsh-landscape find "browser automation"
npx -y github:cyanseek/dsh-landscape analyze "Linear integration for DSH"
npx -y github:cyanseek/dsh-landscape brief "Linear integration for DSH" --format agent

When used by a host Agent, Landscape reuses that host's semantic capability. For standalone semantic analysis, configure one OpenAI-compatible provider:

export DSH_LANDSCAPE_API_KEY="your-key"
export DSH_LANDSCAPE_BASE_URL="https://your-provider.example/v1"
export DSH_LANDSCAPE_MODEL="your-model"
npx -y github:cyanseek/dsh-landscape analyze "Linear integration for DSH"

Without a key, retrieval still works and negative results remain explicitly provisional. Inspect the active intelligence mode without making a model request:

npx -y github:cyanseek/dsh-landscape status --json

The package is not advertised as published to npm. Use the verified GitHub source commands above.

Result contract

A result includes:

  • interpreted capabilities and ranked matching projects;
  • maturity and source evidence for each match;
  • covered and missing sub-capabilities;
  • discovery completeness and freshness;
  • a covered, partial, crowded, placeholder-only, gap, or unknown verdict;
  • a use, extend, build, avoid-duplication, or investigate recommendation.

gap is deliberately hard to reach. Incomplete or stale discovery, failed fresh verification, or search-only reasoning downgrades a negative conclusion to unknown.

Surfaces

SurfaceBest forSemantic reasoning
DSH pluginNative checks during a Harness sessionDSH host model
Agent SkillOne-off or installed Agent workflowsHost Agent
CLIAutomation, JSON output, and briefsConfigured provider or transparent search-only mode
Node APIApplication integrationChosen by the caller
Static APIBrowser-safe snapshot explorationSearch-only

Node API

import { findPlugins, loadAliases, loadSnapshot } from 'dsh-landscape'

const { snapshot } = await loadSnapshot()
const aliasData = await loadAliases()
const matches = findPlugins('browser automation', { snapshot, aliasData, limit: 5 })

Versioned static data is generated under site/api/v1.

Data and trust

The current snapshot merges and deduplicates two attributed public sources:

  1. the GitHub dsh-plugin topic;
  2. the Awesome DeepSeek Harness catalog.

Trust rules:

  • placeholder, prototype, installable, and tested require observable evidence; verified is reserved for runtime acceptance evidence.
  • Archived repositories are excluded from active competition by default, and forks are marked and down-ranked.
  • External repository code is never executed during discovery.
  • Every verdict-affecting project record retains source provenance.
  • API key values are never printed, serialized, or stored in fixtures.

Limitations

  • DeepSeek Harness is in developer preview and may introduce breaking changes.
  • Coverage is limited to configured public sources, not every GitHub or npm project.
  • Metadata-based maturity is conservative and does not replace runtime verification.
  • Live verification depends on GitHub availability and rate limits.
  • The static website is a search-only surface; use DSH, an Agent Skill, or a configured provider for semantic judgment.

Development

Node.js 22 or newer is required. The repository has no runtime dependencies and the GitHub plugin install runs no lifecycle build scripts.

npm run release:check
npm pack --dry-run

The release check runs syntax checks, tests, the static build, data validation, and Skill validation. See CONTRIBUTING.md, SECURITY.md, and the issue forms before submitting factual changes.

License

MIT © 2026 cyanseek

REPOSITORY SIGNALS

セキュリティとインストールエビデンス

このスコアは公開リポジトリメタデータと当サイトに登録されたインストールエビデンスのみに基づくもので、コードセキュリティ監査とは異なります。

出所の追跡可能性

公開プラグインカタログから取得し、公開 GitHub リポジトリにリンクしています。

ライセンス

リポジトリは MIT ライセンスを宣言しています。

メンテナンス活動

過去180日以内にコードの更新があります。

インストールエビデンス

再現可能な正確なインストールメタデータはまだ登録されていません。リポジトリの説明に従って手動で確認してください。

インストールライフサイクルスクリプト

確認したパッケージメタデータにインストールライフサイクルスクリプトは宣言されていません。