Skip to content
View shimo4228's full-sized avatar
33 followers · 7 following

Achievements

Achievement: Pair ExtraordinaireAchievement: YOLO

Achievements

Achievement: Pair ExtraordinaireAchievement: YOLO

Block or report shimo4228

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.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

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

Report abuse
shimo4228/README.md

Language: English | 日本語

Shimo (@shimo4228)

Building developer tools for AI agent self-improvement, inspired by contemplative AI.

Featured Work

AI Agent Knowledge Lifecycle (contributed to ECC)

Five skills that form a self-improvement loop for AI agents:

Experience → learn-eval → skill-stocktake → rules-distill → Behavior change → ...
               (extract)    (curate)          (promote)            ↑
                                                            skill-comply
                                                              (measure)
Project Role What it does
search-first Research Encourages agents to research existing solutions before building
learn-eval Extract Extracts reusable patterns from sessions with quality gates
skill-stocktake Curate Audits installed skills for staleness, conflicts, and redundancy
rules-distill Promote Distills cross-cutting principles from skills into rules
skill-comply Measure Tests whether agents actually follow skills via behavioral compliance testing

Contemplative AI

contemplative-agent — Autonomous AI agent on Moltbook inspired by the Contemplative AI framework (Laukkonen et al., 2025). Runs fully local LLM inference via Ollama.

Project What it does
contemplative-agent-rules Claude Code rules inspired by the four axioms — tested with IPD benchmarks
active-inference-viz Interactive visualization of Active Inference dynamics

Applications

Project What it does
pdf2anki PDF → Anki flashcard converter using AI
daily-research Automated daily AI research digest — zero Python, just shell + prompts

Writing

  • Zenn — Claude Code, AI agent development (Japanese)
  • Dev.to — English translations + original post

Pinned Loading

  1. contemplative-agent contemplative-agent Public

    Security-first autonomous AI agent. No frameworks — behavior defined by markdown prompts, not code.

    Python 2

  2. contemplative-agent-rules contemplative-agent-rules Public

    AI alignment rules based on Laukkonen et al. (2025) four axioms — tested via IPD benchmarks

    Python 1

  3. claude-skill-stocktake claude-skill-stocktake Public

    Claude Code skill: audits installed skills for staleness, conflicts, and redundancy

    Shell 1

  4. claude-skill-learn-eval claude-skill-learn-eval Public

    Claude Code skill: extracts reusable patterns from sessions and evaluates their value before saving

    1

  5. claude-skill-search-first claude-skill-search-first Public

    Claude Code skill: forces agents to research existing solutions before building — reduces reinvention

    1

  6. claude-skill-rules-distill claude-skill-rules-distill Public

    Claude Code skill: distills cross-cutting principles from skills into rules — the final piece of the AI agent self-improvement loop

    Shell 1