By zoahdev · 2026-08-15 · All output below is real GitHub API data
dsh plugin --profile web add github:zoahdev/dsh-github-intelligence
Restart dsh web and just ask. No API key needed (60 requests/hour anonymous; the plugin's 60s TTL cache makes that budget go far).
deepseek-ai/deepseek-harness."ollama/ollama have right now?"This is the real github_repo_report output for deepseek-ai/deepseek-harness on 2026-08-15:
# deepseek-ai/deepseek-harness
deepseek-ai/deepseek-harness — DeepSeek Harness: Everything is a Plugin.
Stars: 98472 · Forks: 9214 · Open issues: 0 · Default branch: master
Language: TypeScript · License: MIT · Pushed: 2026-08-13
https://github.com/deepseek-ai/deepseek-harness
Latest release: none
Open issues: 0
Recent commits:
- 47f9438 Merge pull request #2519 from deepseek-harness/feat/npm-public (imccyu)
- abe560f release(dsh): 0.1.0-rc.5 (imccyu)
- 8c1e8d9 build(release): publish the dsh family publicly (imccyu)
- f26a6f6 Merge pull request #2520 from deepseek-harness/docs/paper (Tianyi Cui)
- 124aa5f Merge pull request #2521 from deepseek-harness/release/dsh-0.1.0-rc.3 (imccyu)
Top contributors: 1. tianyicui (5235) · 2. LegGasai (1361) · 3. imccyu (1168) · 4. Chinesezjc (587) · 5. turtle1999 (585)
One question = full overview + latest release + open issues + recent commits + top contributors. That is due diligence.
| Tool | What it does |
|---|---|
github_repo |
Full overview (topics, default branch, archived status) |
github_releases |
Recent releases with body previews |
github_issues |
Issues by state |
github_pulls |
Pull requests by state (merge status included) |
github_contributors |
Top contributors by commit count |
github_search |
Repository search by stars or last update |
github_repo_report |
All of the above in one canonical answer |
The anonymous GitHub API allows 60 requests/hour. Every endpoint in this plugin has a 60s TTL cache: ask for the same repo report three times in a row and only the first costs 4 requests. The deep report runs its sub-calls in parallel and reuses the cache instead of stacking cost.
GitHub's /issues endpoint returns pull requests by default. The plugin filters them out by the pull_request field in listIssues — this is the difference between "integrating" and "calling an API": same surface, clean data.