ruler770525/dsh-anchored-flash
Anchored-standard agent preset for DeepSeek Harness: Minimal-aligned first request, low-injection post-promotion, indirect AGENTS.md loading — anchored intelligence without the IQ drop. （DeepSeek Harness 的锚定标准预设：首轮 Minimal 对齐、晋升后低注入、AGENTS.md 间接按需加载——锚定智力不降智。）
安装Install
dsh plugin --profile web add github:ruler770525/dsh-anchored-flash把 ruler770525/dsh-anchored-flash 加入你的 DSH 配置（web profile）即可启用。
READMEREADME
dsh-anchored-flash
English | 中文
An experimental DeepSeek Harness agent preset: Minimal-aligned anchoring with low-injection post-promotion, an indirect AGENTS.md injection pattern, and subagent anchoring — the "anchored intelligence without the IQ drop" experiment series.
Fork of xiaobright/dsh-anchored-standard
with local modifications; the experiment report is in
upstream issue #49.
Maintenance status (2026-08-17): upstream
xiaobright/dsh-anchored-standardhas entered maintenance mode (FAREWELL — the official API price change made the experiment series cost-prohibitive). This repository, an experiment branch of that series, is frozen: it no longer tracks upstream, and the experiment conclusions live in upstream issue #49. Related work has moved to thezhu1090093659/dsh-web-uiecosystem (the non-assertive AGENTS.md hint,instructionHint, was merged there as #388). Local installs keep working; no further releases are planned here.
What makes this different from upstream
| Feature | Upstream anchored-standard |
This preset |
|---|---|---|
| First request | Minimal pair bash + str_replace_editor, maxTokens 1024, zero injected context |
same |
| System prompt after promotion | complete: true persona, stays the Minimal sentence (46 chars) forever |
same, enforced by the assemble filter in every phase (tool-bootstrap) |
| Post-promotion catalog | bootstrap pair + discovery tools + dev_tool_search unlocks |
configurable residentTools whitelist (default: 18 tools incl. memory/dtodo/web_search/de_session/read_image/exit_plan_mode/skill_search/skill_load) |
instruction-hint wording |
imperative ("read the relevant instruction files first and follow them") — measured to flip the anchored trajectory back to "let me" | neutral/suggestive reference note — measured to preserve "we" (see experiment table below) |
| Subagents | opt-in includeSubagents |
default true — subagents go through the same anchor flow |
| AGENTS.md | full digest injection (perturbs) or nothing | indirect injection: 1 KB topic index + AGENTS-*.md topic files + env-* skills, read on demand (skill_load / read) — measured non-perturbing |
| Windows bash | custom-bash (fixed path) |
custom-bash with PATH fallback when the configured Git Bash path is missing |
| Runtime context (memory snapshot etc.) | stripped | stripped in every phase (low-injection) |
Experiment summary (measured 2026-08-16, DeepSeek V4 Pro, reasoningEffort=max)
Fingerprint = completed-reasoning counts of we / let's / let me
(case-insensitive, whole words). Sessions are real DSH sessions on Windows +
rc.6.
| Exp | Hint wording | System prompt | Result |
|---|---|---|---|
| baseline | imperative ("read first and follow them") | restored to standard (46→6620 chars) | we→let me flip after promotion (0 we / 3 let me; model: "I must read relevant instructions") |
| E1 | neutral reference note | 46 chars in every phase | we preserved (up to 49 we / 1 let me in 22.5k-char analyses) |
| E2 | neutral + user asked to read AGENTS.md | 46 chars | full on-demand chain walked (skill_search → skill_load → read 1 KB index); we preserved; environment facts actually used |
| E1.5 | suggestive ("reading the index is recommended") | 46 chars | we preserved; model reads only when the task needs environment facts |
Key conclusions:
- The promoted system-prompt mutation (46→6620 chars) is itself a perturbation. Keep the system prompt at the single Minimal sentence in every phase.
- Imperative wording in a user-role injected message correlates with the style switch (we report correlation, not mechanism).
- Non-imperative reference wording does not flip the trajectory, and read willingness is driven by task need, not hint wording.
- Anchored intelligence and AGENTS.md availability are not mutually exclusive: tell the model (non-imperatively) that reference documents exist, and let it load them on demand.
Install
$target = Join-Path $env:USERPROFILE '.dsh\.agent-presets\anchored-flash'
if (Test-Path -LiteralPath $target) { throw "Preset already exists: $target" }
New-Item -ItemType Directory -Force -Path (Split-Path -Parent $target) | Out-Null
Copy-Item -Recurse -LiteralPath '.\preset' -Destination $target
Restart DeepSeek Harness, create a fresh session, pick 锚定标准·满血子代理
(anchored-flash). Do not switch presets mid-session.
Recommended AGENTS.md layout (the indirect-injection pattern)
Split the user-global ~/.dsh/AGENTS.md into:
~/.dsh/AGENTS.md— ~1 KB topic index (3 iron rules + pointer table)~/.dsh/AGENTS-<topic>.md— topic files (environment / network / dsh / workflow / browser …)~/.dsh/skills/env-*— the same content as skills, discoverable viaskill_searchand loadable viaskill_load
Verify
node verify/verify-anchored-flash.mjs # 18 logic checks, no DSH runtime needed
node verify/trace-session.mjs <session-dir> # header + reasoning fingerprint trace
node verify/count-pronouns.mjs <session-dir> # we / let's / let me counters
node verify/dive-session.mjs <session-dir> # full system + message sources
node verify/dump-session.mjs <session-dir> # header + first-line dump
License
MIT. The preset is based on the DeepSeek Harness Standard preset and on
xiaobright/dsh-anchored-standard; see NOTICE and
LICENSE. Part of the experiment design and this documentation
were assisted by DeepSeek (an AI assistant); all measurements are from real
DSH sessions.
相关推荐Related
- deepseek-ai/deepseek-harness DeepSeek Harness: Everything is a Plugin.
- deepseek-ai/deepseek-harness DeepSeek Harness (dsh) — 由 DeepSeek AI 开发的开源 agent harness（智能体框架）。一切皆插件：核心 + 插件 + 技能 = 可热插拔的 AI 编程工作台。DeepSeek Harness: …
-
ruvnet/ruflo
🌊 The original agent meta-harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build c…
-
amruthpillai/reactive-resume
A one-of-a-kind resume builder that keeps your privacy in mind. Completely secure, customizable, portable, open-source a…
-
volcengine/OpenViking
Self-evolving Context Database for AI Agents. Unify Agent Memory, Knowledge RAG and Skills.
-
Molunerfinn/PicGo
:rocket: The Ultimate Image Uploader for Efficient Creators. Supports Obsidian, Typora, VS Code etc. and 60+ image hosti…
-
titanwings/colleague-skill
将冰冷的离别化为温暖的 Skill，欢迎加入数字生命1.0！Transforming cold farewells into warm skills? It's giving rebirth era. Welcome to Digital …
-
nocobase/nocobase
NocoBase is an open-source AI + no-code platform for building business systems fast. Instead of generating everything fr…