deepseek-ai / deepseek-harness
DeepSeek Harness: Everything is a Plugin.
github.com/deepseek-ai/deepseek-harnessIndexed at 4878cdaUp to date with upstream
About this codebase
DeepSeek Harness is a developer-preview harness that ingests configuration and repository/session data, runs DeepSeek-oriented agent workflows through its CLI/SDK and sandboxed execution layers, and produces session artifacts plus API-served and UI-presented results for developers building and evaluating DeepSeek integrations. Inputs typically include harness/run configuration, prompt/session state, and workflow definitions; the pipeline builds a usable runtime (including preset composition), executes the workflow/turns via the command/runtime/sandbox executors, and emits durable session outputs (including checkpoints and projected metadata) that downstream tools can query and display.
Written by gpt-5.4-nano from the index
12,578 files and 39,595 symbols in 11 modules, led by TypeScript (5,108 files) and JavaScript (109).
Start reading
Code health
This codebase scores 6.6 out of 10 for code health, which we rate fair. It also scores maintainability 8.1 and static performance 9.8 out of 10. The three are scored separately and never blended into one number. Risk is concentrated, as it usually is: 1,306 of 12,578 files are git hotspots, and they average 5.4 — which is where the fixes pay off most.
On the leaderboards:#4 of 6 Coding agents#27 of 37 TypeScript repos
Do next
4 things worth doing in the week to Sep 28, the last indexed commit, 4 of them now.
Now
- Critical
Clean up what this week's commits left: 6 serious findings in 5 files
4 commits in the last 7 days added or worsened them, and they are still open. Start with
packages/core/agent-loop/src/agent.ts, which has 2 critical.- serious findings
- 6
- critical
- 3
- files
- 5
Medium effort - High
Start breaking up ModelSelect (CCN 103, 405 lines): first lift lines 101-106
packages/client/ui-model-selection/src/client/ModelSelect.tsx:101
ModelSelect is long, branchy and widely depended on; 2 files import it, changed 20 times in 90 days.
- gain
- +1.0 health on this file(inferred)
- health recoverable
- +1.0(inferred)
- size
- CCN 103, 405 lines, nests 3 deep
- steps
- 1 (1 mechanical)
Small effort - High
Start breaking up SidebarRoot (CCN 41, 183 lines): first lift lines 122-132
packages/client/ui-sidebar/src/client/SidebarRoot.tsx:122
SidebarRoot is long, branchy and widely depended on, and part of it is duplicated, so splitting it also removes a copy; 3 files import it, changed 41 times in 90 days.
- gain
- +1.0 health on this file(inferred)
- health recoverable
- +1.0(inferred)
- size
- CCN 41, 183 lines
- steps
- 1 (1 mechanical)
Small effort
Use deepseek-ai/deepseek-harness in your agent
Point Claude Code, Cursor, Codex or VS Code at this repository. Your agent gets the index this page is built from: cited answers, callers, history and health for any file.
https://api.repowise.dev/mcp/deepseek-ai/deepseek-harnessOne command, run anywhere. Adds the server to your local Claude Code config.
Docsclaude mcp add --transport http repowise https://api.repowise.dev/mcp/deepseek-ai/deepseek-harness \ --header "Authorization: Bearer rw_live_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
This repo is public, so anyone with a Repowise account and an API key can connect.
Try asking
- get_overview()
- get_answer("how is deepseek-harness put together?")
- get_why("why is deepseek-harness structured this way?")
Prefer local? pip install repowise && repowise init indexes your own checkout, no account needed.
Go deeper
- OverviewEverything the index found, what changed this week, commits and decisions.
- DocumentationModule-by-module documentation, 4,559 pages, 83 written by a model.
- ArchitectureDependency graph, layers and third-party dependencies.
- Code healthEvery file's score, the findings behind it, tests and refactoring plans.
- CommitsEach commit's effect on code health, with agent provenance.
- ContributorsWho owns what, and where one person holds most of the knowledge.
- DecisionsArchitectural decisions mined from commits and pull requests.
- Ask this codebaseSign inGrounded answers over the indexed code.
Maintain this repo? Claim the page
Sign in with GitHub and we check that you are an admin or maintainer of deepseek-ai/deepseek-harness. Then add a badge to the README that links readers here.
Use this repo from Claude or ChatGPT
Add this URL as a custom connector and sign in with repowise when asked. No API key to copy.
https://api.repowise.dev/mcp/deepseek-ai/deepseek-harnessFor maintainers
Add a badge to your README
Every badge is public, cached, and updates on its own after every index. Nothing to install.
Repowise badge
Links to this page, with no score on it. Adding it also re-indexes the repo every week.
[](https://repowise.dev/repo/deepseek-ai/deepseek-harness?src=badge_wiki)Code health badge
The repo's health score out of 10, from the latest index.
[](https://repowise.dev/repo/deepseek-ai/deepseek-harness?src=badge_health)Code health card
A larger card with the score, its change since the last index, and when it was measured. The HTML version follows the reader's light or dark theme.
[](https://repowise.dev/repo/deepseek-ai/deepseek-harness?src=badge_card)How deepseek-ai/deepseek-harness works
DeepSeek Harness: Everything is a Plugin. This page is a map of the deepseek-ai/deepseek-harness repository, written primarily in TypeScript, rebuilt from the source each time it is indexed. Repowise parses every symbol, computes a dependency graph, scores per-file code health from complexity, duplication, test coverage and churn, mines git history for hotspots and ownership, and lifts the architectural decisions into documentation you can read here or query through MCP.
The codebase has 12,578 files and 39,595 symbols in 11 modules, led by TypeScript, JavaScript and Python. Code health is 6.6 out of 10, rated fair.
Use the links above to open each view, or connect this repository to your agent for grounded answers inside Claude, Cursor, Codex or VS Code.