dsh-tacit
Learns what you leave unsaid in your prompts and steers the DeepSeek Harness agent for you
- Stars
- 0
- Language
- JavaScript
- Created
- Aug 27, 2026
- Updated
- Aug 27, 2026
Introduction
Tacit
Learns what you leave unsaid in your prompts — and tells the agent for you.
A plugin for DeepSeek Harness.
You keep writing prompts the way you already do. Tacit watches how each turn actually went — retries, tool errors, and above all the message you send when the agent got it wrong — and turns your habits into a few directives the agent follows in every new session. No clicks, about $0.001 per lesson.
you: "make the login page better"
agent: …stumbles…
you: "no, I meant the Next.js app under apps/web"
Tacit: learns → "The user often omits which app they mean — check apps/web first."
injects it into every later session's system prompt.
You never have to say it again.
Install — 30 seconds
dsh plugin --profile web add dsh-tacit
Restart dsh web, refresh the page. Optional kick-start: Settings → Tacit →
Learn from my last 20 turns (≈ $0.02, once).
Requires DeepSeek Harness >= 0.1.1-rc.1, Node >= 22, and a DeepSeek API key
already configured in the harness (Tacit never reads it).
What you get
- Zero-click learning — messy turns and your own corrections are analyzed in the background, with the previous turn as context. Capped per day (30 by default). ≈ $0.001 per analysis.
- Directives that earn their place — learned directives are injected as a small system-prompt section you can read, edit, toggle or delete. A new one starts as a candidate and is retired automatically if your messy-turn rate gets worse. free
- ✨ Improve — a composer button that rewrites your current draft using what Tacit has learned, with a before/after preview and 👍/👎. ≈ $0.001 per click
- Measured, not guessed — Settings shows your real trend: messy-turn rate and tokens per turn, first 20 turns vs. latest 20. free
How it works
- A pure fold over the session's events produces a bounded digest per turn.
- When a turn ends messy, or your next message reads as a correction, one
small
deepseek-v4-flashcall (low reasoning effort, structured output) analyzes the prompt. - Every few analyses, one more call distills the findings into 2–4 directives.
- The directives are rendered as a ≤300-token section of the system prompt — the exact text is visible in Settings, and nothing else about your prompts leaves the machine.
Privacy & cost, in one paragraph
Tacit never sees your API key (every call goes through the harness's own model
service), only calls the allowlisted official model, keeps reports and
directives in ~/.dsh/storages/tacit/, refuses cross-site requests to its own
routes, and never deletes anything but its own reports. The dollar figures are
estimates at list price; a cost plugin such as dsh-cost-meter shows the real
number. Full details, including the honest list of limitations:
docs/privacy-and-cost.md.
More
Configuration · Privacy, cost & limitations · Architecture · Contributing · Changelog
MIT © hackernotfound