Skip to content
View alloevil's full-sized avatar
:electron:
Working from home
:electron:
Working from home

Block or report alloevil

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
alloevil/README.md

Typing SVG


🧭 What I Do

Every tool here exists because an LLM told me something that turned out to be false — a diagram that was the model's opinion, a judge score that could be bluffed, a "refactor" that dropped a symbol, a benchmark number with no source. So each one replaces a claim with a check.

   

Impact Analysis · Deterministic code graphs, mutation-tested recall, file:line on every edge
Honest Evaluation · Program gates before rubrics, paired statistics, verdicts that say what they can't rule out
Agent Observability · Read what the agent actually did — turn by turn, token by token


🔭 Featured Projects

Full index with every project, grouped by area: alloevil.github.io/projects

💥 codeblast

Know what breaks before you merge — deterministic code graph with impact, change & architecture maps. Recall verified by mutation testing (tRPC, 950 files: 28/28).

npx codeblast demo   typescript impact-analysis agent-skill

→ Explore · Live demo

📐 paired-eval

Evaluate models, agents and harnesses: program checks first, rubrics for the rest, honest paired statistics. Zero dependencies.

pip install paired-eval   python evaluation statistics

→ Explore · Docs

🔬 AgentXRay

Web dashboard for AI agent session logs — Claude Code, Codex, OpenClaw, Hermes, OMP, Gemini CLI. Per-turn time / token / cost ledger.

npx @alloevil/agent-xray   javascript observability debugging

→ Explore

🔍 coding-agent-internals

How 12 coding agents implement search, edit, LSP, DAP and sub-agents — shell-fork vs in-process, str_replace vs hash-anchored. Ceilings, not checklists.

research claude-code codex omp

→ Explore

🧾 agents-with-receipts

Evidence-first agentic coding: a cell-by-cell verified cross-tool table, a sourced practices map, and an AGENTS.md linter.

practices agents-md linter

→ Explore

🧪 deepresearch-arms-lab

14-arm ablation of deep research pipelines on a weak base model — negative results left in, n=15 re-validation.

ablation-study llm-evaluation negative-results

→ Explore


📚 Recently Reading

🛠️ Tech Stack


📊 Evidence, Not Widgets

These cards are rendered daily from the tools' own outputs — a real repo's dependency graph, the latest mutation-testing run, a real session's per-turn ledger. Same numbers as the tools print.

codeblast architecture map — module dependencies, circular dependencies in red codeblast mutation-tested recall

AgentXRay per-turn ledger — time and tokens per step

github contribution grid snake animation

coding cat

Code Time

I'm an Early 🐤

🌞 Morning                1805 commits        █████████████░░░░░░░░░░░░   52.95 % 
🌆 Daytime                1317 commits        ██████████░░░░░░░░░░░░░░░   38.63 % 
🌃 Evening                275 commits         ██░░░░░░░░░░░░░░░░░░░░░░░   08.07 % 
🌙 Night                  12 commits          ░░░░░░░░░░░░░░░░░░░░░░░░░   00.35 % 

📅 I'm Most Productive on Monday

Monday                   1603 commits        ████████████░░░░░░░░░░░░░   47.02 % 
Tuesday                  230 commits         ██░░░░░░░░░░░░░░░░░░░░░░░   06.75 % 
Wednesday                183 commits         █░░░░░░░░░░░░░░░░░░░░░░░░   05.37 % 
Thursday                 527 commits         ████░░░░░░░░░░░░░░░░░░░░░   15.46 % 
Friday                   188 commits         █░░░░░░░░░░░░░░░░░░░░░░░░   05.51 % 
Saturday                 58 commits          ░░░░░░░░░░░░░░░░░░░░░░░░░   01.70 % 
Sunday                   620 commits         █████░░░░░░░░░░░░░░░░░░░░   18.19 % 

📊 This Week I Spent My Time On

🕑︎ Time Zone: Asia/Shanghai

💬 Programming Languages: 
Python                   7 hrs 40 mins       ████████░░░░░░░░░░░░░░░░░   33.70 % 
Markdown                 5 hrs 31 mins       ██████░░░░░░░░░░░░░░░░░░░   24.22 % 
JavaScript               3 hrs 20 mins       ████░░░░░░░░░░░░░░░░░░░░░   14.65 % 
TypeScript               2 hrs 6 mins        ██░░░░░░░░░░░░░░░░░░░░░░░   09.24 % 
Other                    1 hr 4 mins         █░░░░░░░░░░░░░░░░░░░░░░░░   04.75 % 

🔥 Editors: 
OMP                      18 hrs 15 mins      ████████████████████░░░░░   80.11 % 
Cursor                   2 hrs 34 mins       ███░░░░░░░░░░░░░░░░░░░░░░   11.32 % 
Claude Code              1 hr 47 mins        ██░░░░░░░░░░░░░░░░░░░░░░░   07.84 % 
Unknown Editor           8 mins              ░░░░░░░░░░░░░░░░░░░░░░░░░   00.59 % 
Agent                    1 min               ░░░░░░░░░░░░░░░░░░░░░░░░░   00.14 % 

🐱‍💻 Projects: 
isc-world-model          5 hrs 21 mins       ██████░░░░░░░░░░░░░░░░░░░   23.54 % 
wca                      2 hrs 54 mins       ███░░░░░░░░░░░░░░░░░░░░░░   12.78 % 
codeblast                2 hrs 37 mins       ███░░░░░░░░░░░░░░░░░░░░░░   11.54 % 
chainsentry-ai           2 hrs 3 mins        ██░░░░░░░░░░░░░░░░░░░░░░░   09.06 % 
AgentXRay                1 hr 25 mins        ██░░░░░░░░░░░░░░░░░░░░░░░   06.24 % 

💻 Operating System: 
Linux                    22 hrs 47 mins      █████████████████████████   100.00 % 

I Mostly Code in Python

Python                   14 repos            █████████░░░░░░░░░░░░░░░░   34.15 % 
JavaScript               10 repos            ██████░░░░░░░░░░░░░░░░░░░   24.39 % 
TypeScript               9 repos             █████░░░░░░░░░░░░░░░░░░░░   21.95 % 
HTML                     3 repos             ██░░░░░░░░░░░░░░░░░░░░░░░   07.32 % 
SCSS                     1 repo              █░░░░░░░░░░░░░░░░░░░░░░░░   02.44 % 

Last Updated on 09/09/2026 02:13:18 UTC


Profile Views

Pinned Loading

  1. weibo-chat-auto weibo-chat-auto Public

    微博群聊消息自动归档 · 原生桌面应用(应用内扫码登录)+ 本地可视化查看器,含 AI 每日摘要与 Agentic 问答

    JavaScript 19 7

  2. AgentXRay AgentXRay Public template

    Web dashboard for viewing AI agent session logs — supports OpenClaw, Codex, and Claude Code | 多平台 AI Agent 日志查看面板

    JavaScript 16 2

  3. github-discovery github-discovery Public

    🔥 Discover trending GitHub repos before they go mainstream. 6 data sources · smart scoring · anti-spam · daily email digest.

    Python 2

  4. NousResearch/hermes-agent NousResearch/hermes-agent Public

    The agent that grows with you

    Python 244k 50.3k