perrylink/dsh-data-quality
Deterministic data profiling, cleaning and verification: data_profile / data_clean / data_verify three model tools + frozen cross-plugin verifyCitations reference verification contract, report persistence to data_quality storage domain; npm dsh-data-quality has been released 0.1.3
Deterministic data profiling, cleaning and verification for DeepSeek Harness. All computation is plain TypeScript inside the harness process — the model never does the math. Three model tools (profile, clean, verify) sit on a ctx.dataQuality service seam that other plugins can consume, including a frozen verifyCitations(request) contract for checking numbers and strings cited in a document against a dataset snapshot. Every run persists a durable report (JSON) to the data_quality storage domain, keyed by run timestamp plus a dataset-path fingerprint.
Install
dsh plugin --profile web add dsh-data-qualitynpm package dsh-data-quality 0.3.9 (registry-verified 2026-09-10; registry repository field -> github.com/PerryLink/dsh-data-quality). The dsh plugin command requires pnpm on PATH. Restart dsh web after installing (bundles activate on restart). Uninstall: dsh plugin --profile web remove dsh-data-quality.
Compatibility
DeepSeek Harness dsh-v0.1.5-rc.1 (README adapted 2026-09-10; verified against the v0.1.5-rc.1 master checkout). Node.js ^22.19.0 || >=24.0.0. pnpm@11.7.0. Host-only plugin (Windows / macOS / Linux).
Details
- Repo: perrylink/dsh-data-quality
- Category: Files & Data
- Stars: 11
- Version: npm dsh-data-quality 0.3.9 (registry-verified 2026-09-10)
- Last push: 2026-09-03
- First seen: 2026-08-19
Recent updates
Adapted to DeepSeek Harness dsh-v0.1.5-rc.1 (2026-09-10): the session envelope keeps its ignorable field for stored-log read compatibility only — Session.append still cannot stamp it, so audit-gate behavior is unchanged.
FAQ
- How do I install dsh-data-quality?
- Run: dsh plugin --profile web add dsh-data-quality (npm channel). The dsh plugin command needs pnpm on PATH. A git channel is also documented, plus a tarball channel via pnpm pack. Restart dsh web afterwards.
- What tools does dsh-data-quality add?
- Three dataset operations — profile, clean and verify — plus data_report to read persisted reports back from the data_quality storage domain by exact reportKey or by kind.
- Where are data-quality reports stored?
- Every profile/clean/verify/citation run persists to the data_quality storage domain (JSON backend), keyed by run timestamp plus a dataset-path fingerprint.
Alternatives
imsai-sh/zhuzhiliao · bruc3van/awesome-dsh-plugin · ZSeven-W/dsh-openpencil