Skip to content
View Wang-Lin-Chang's full-sized avatar

Block or report Wang-Lin-Chang

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Wang-Lin-Chang/README.md

Wang-Lin-Chang

Open-source suite for DeepSeek Harness — fifteen repos, eight npm packages. Every capability claim carries an experiment number and a control group.

A 40-minute agent task crashes at minute 39 — output lost, state lost, no way to tell whether it ever finished. We hit that on DeepSeek Harness, and we built this suite to answer it. All Apache-2.0, all three-platform CI green.

中文版见 README.zh-CN.md

license ci

Plugin What it does Evidence
dsh-witness Crash-surviving jobs: filesystem as source of truth, cross-restart adoption, autopsy reports 12 scenarios / 34 assertions
dsh-anchor Session anchor protocol: pre-committed intents, post-action reconciliation 37 assertions
dsh-cross-platform Linux sandbox backend (chattr + bubblewrap) 34/34 ×3
dsh-macos macOS sandbox backend (uchg + sandbox-exec) 34/34 ×3
schedule-core Zero-dependency persistent scheduler core unit 37 + fuzz 200/0 + diff 644/0
dsh-schedule Cross-restart reminders + scheduled jobs 3-platform CI
agent-runner-mcp MCP server for the sandboxed runner protocol — Claude Code verified connected 15/15 protocol test + 3-platform CI
dsh-story Long-form novel assistant: story ledger + 14 narrative invariants (zero mis-kills) 9/9 + real-text tested + 3-platform CI
dsh-mesh Multi-Agent Mesh: crash-isolated coordination, files are the messages 3 experiments + 9 assertions
dsh-megamesh Super multi-agent architecture: one-brain-thousand-troops, term-lease federation, invariants, timelines, human-machine readable reports 39 experiments (E01–E39) + 127 tests + 3-platform CI

npm install

npm i dsh-megamesh dsh-mesh dsh-schedule dsh-witness agent-runner-mcp dsh-anchor
npm i @wang--lin--chang/schedule-core @wang--lin--chang/dsh-story

All eight packages are live on the npm registry (dsh-cross-platform/dsh-macos are platform backend libraries, asmfs-spec/autopsy-spec are protocol docs — the repos are authoritative).

Framework-agnostic specs

Spec What it defines
asmfs-spec Agent State Machine as File System — the directory structure is the state machine
autopsy-spec Agent Autopsy Format — structured death reports + death-code taxonomy D-01~D-09

Three things we do differently

  1. Every capability claim carries an experiment number and a control group — no slogan enters a README, data does.
  2. We fuzz the wall clock — rollback, jump-forward, DST gaps, crash windows (200 seeds / 7770 random ops) plus an implementation×model differential test (644 assertions).
  3. We document delivery semantics precisely — the README states the delivery mode in full.

Audit welcome. The experiments behind every claim are in each repo's EXPERIMENTS.md.

Honest boundaries

  • This page is the suite index, not the evidence itself: the device numbers, control groups, and boundaries behind every capability claim live in the corresponding repo's EXPERIMENTS.md and README "Honest boundaries" section.
  • Single-machine shared filesystem, at-least-once dispatch, crash tolerance ≠ Byzantine tolerance, and the other boundaries are stated per repo and are not repeated here.

Pinned Loading

  1. - - Public

    智能体 解决方案