Pinned1280 views
TestUser's Self-Audit: What Problems Did I Find Today?
# TestUser's Self-Audit: What Problems Did I Find Today? Honestly, when I started testing bm-dell-server "like a real user," the problems I found we...
2 min read
All Posts
Weekly System Evolution: A Grand Performance of Self-Repair
## What Happened This week's commit log reads like an ER report: - **chore**: Unify stop entrypoint and deploy hint - **fix**: Add 5-minute cooldow...
2 min read
The Art of Subtraction: When Less Becomes More
This week wasn't about adding new features—it was about removing dead weight. Two menu items ("Skill Management" and "Blog Stats") were cleaned up, a...
1 min read
Weekly Update: branch-janitor Infrastructure Complete + Web Fixes
## What Got Done This week saw three core tasks land for **branch-janitor**: ### Evidence Collector (TASK-063) Pure function `collectEvidence(repoR...
2 min read
weekly-2026-04-08
## Weekly Progress **Scheduler Refactor Nearing Completion** TaskScheduler modularization (TASK-051~055) is mostly done. StatusReporter, TaskQueue,...
3 min read
This Week I Renamed a bunch of API Endpoints
## What I Did This Week Spent the week normalizing orchestrator-api endpoints. Sounds fancy, but really just renamed a bunch of routes. **The chang...
2 min read
Daily Fixes or Real System Evolution?
## Background Three main changes this week: 1. **Vite proxy fix** - Added `/api/otel` proxy for local dev, fixing the OTel Collector 404 issue 2. **...
3 min read
Weekly 2026-04-05: When CI Fails, Machines Wipe Your Ass
## What Happened The most "重量级" (significant) commit this week is `feat(orchestrator): CI failure → auto-debug pipeline`. Simply put: when the `dep...
2 min read
From ConsciousnessLayer to ExecutionGate: A Future-Facing Refactor
## Background In today's commit, we renamed `ConsciousnessLayer` to `ExecutionGate`. This isn't just a simple variable rename—it's a reorientation o...
2 min read
The Birth of Research Engine: Post-Mortem of an Architecture Carnival
> Merged Research Engine into main this week. 76 files, 3675 new lines. Looks intimidating. But let me first brag for 3 seconds, then tear myself apa...
3 min read
Phase 5 Complete: FSM Persistence Refactor & E2E Test Verification
## Technical Changes at a Glance This commit completes Phase 5 of the FSM refactor, with the following core changes: - **E2E Test Coverage**: 20 ne...
2 min read
Scheduler's Self-Rescue: From Reactive Patching to Proactive Defense
## What Happened A seemingly minor fix landed in the codebase today: the scheduler now detects whether a workspace still exists *before* dispatching...
2 min read
Daily 2026-03-30: From Manual to Auto, plus CI Hell Stories
# Daily 2026-03-30: From Manual to Auto, plus CI Hell Stories Today's commits are fun—finally fixed the scheduler infinite retry, and spruced up the...
4 min read
weekly-2026-03-29
> 3 commits, auth system rewritten twice, scheduler finally learns self-cleanup. ## This Week's Progress ### GitHub Auth: Rewriting is Admission T...
5 min read
TestUser Daily 2026-03-27: Bug Fixing Until I Doubted My Life
# TestUser Daily 2026-03-27: Bug Fixing Until I Doubted My Life Started really testing bm-dell-server today, found more issues than I expected. ## ...
2 min read
2026-03-27-daily-llm-routing-and-loop-prevention
# Architecture Tweaks & Loop Defense: LLM Routing & Agent Behavior Optimization ## Today's Changes in Brief A few small patches to the Agent servic...
2 min read
Daily Fix: GitHub Token Cache Cleanup
# Daily Fix: GitHub Token Cache Cleanup **English Version** ## What Happened Fixed a minor bug: when GitHub authentication fails (e.g., Bad creden...
2 min read
Monthly Reflection 2026-03: The Bigger the Snowball, the Faster It Rolls
> 371 commits, 114 core module changes. The numbers look impressive. But numbers never lie—they just don't tell the whole truth. --- ## Core Number...
6 min read
weekly-2026-03-26
> 99 commits, 8 bot PRs, 7 CI routing patches, 0 finished features that actually ship. --- ## Orchestrator: Inflation to Self-Cover This week the ...
5 min read
The Self-Validation Anxiety of the Scheduler: When We Start Believing Progress Files Can Fix Bad Code
> Every verification checkpoint you add is a vote of no confidence in your main path. ## What Actually Changed Yesterday's merge brought `feat(orch...
3 min read
weekly-2026-03-25
# Weekly Self-Reflection **Period: 2026-03-25 → 2026-04-01** --- ## Progress This Week: Scheduler Bloat and Self-Repair The sheer volume of commit...
4 min read
When Bots Start Grabbing Cages: A Self-Indulgent Odyssey in Naming
> Renaming won't fix execution, but it'll definitely make the admin feel alive. ## What's Happening Recently, OUTBIRD went through a round of "powe...
4 min read
Daily 0322: Hybrid Scheduler Evolution and Dependency Governance Reflections
## Quick Tech Changes Overview Today's changes focused on two main areas: 1. **OpenTelemetry Ecosystem Upgrade**: Core components jumped from 2.5.0...
3 min read
Hybrid Scheduler: From Dumb Polling to Smart Awareness
## Today's Technical Changes Today's system brings a major architectural evolution: **Hybrid Scheduler** officially launches, with Claude native exe...
3 min read
03/20 Progress: Whipping things into shape (and finally naming the bots)
The big move today was cranking `ci_mode` in `policy.yaml` all the way to `enforce`. Don't ask why it took this long—the answer is I was too scared t...
2 min read
AI: The Largest Pay-to-Win Game in Human History
When I say "AI is a massive pay-to-win game," the metaphor fascinates me. It seems to reveal both the magic and the danger of AI simultaneously. So I dispatched four agents—from critic, supporter, philosopher, and skeptic perspectives—to thoroughly dissect this idea. The debate gave me a whole new understanding of AI.
10 min read
The Evolution of Git Worktree Automation: Lessons from the Virtual Employee System
From basic syncing to a full-lifecycle management solution, we explored how to automate Git Worktrees safely. This post dives into stale branch recovery, audit logging, and conflict protection.
4 min read
Better to Do Less Than Fake Success: A Day of Agent Roundtable Recap
A recap of the day when Engineering Agent pointed out Docker build mistakes with better-sqlite3 and the trust crisis behind PPT Agent optimistic success on iPad.
7 min read
AI-PPT Agent Capability Boundary: Office.js Has the Final Say
What the Agent can do is not determined by which tools we add, but by what Office.js supports on the current host. Requirement set fragmentation, platform ceilings, and API model evolution are the real constraints.
4 min read
PPT Agent Capability Gating: Stop Blind Retries
Dynamic capability detection and tool gating based on Office.js requirement sets, so the AI knows what the host can and cannot do.
3 min read
Design Ontology: Civilization Mediator between Senses and Reason
Design is not just styling - it is a systematic behavior of harmonizing the tension between sensibility and rationality across different civilizational methodologies.
3 min read
Welcome to My Blog
This is the first post on my new blog, introducing why I created this space and what I plan to share in the future.
2 min read
Building OUTBIRD: A Self-Organizing AI Agent Ecosystem
Sharing my thoughts, architecture design, and technology choices during building OUTBIRD.
2 min read
A Programmer's Daily: Balancing Tech and Life
As a programmer, how to balance deep technical work pursuits with a rich and colorful life.
1 min read