Liam-Skills
The Taixu five-turns skill pack: ideas pass divergence, convergence, verification, critique and a rescue of rejected options.
- 113 ★ on the repository
- Last commit
Minimal access
Why this level
- The skills steer an agent's reasoning and some of them run shell commands to test hypotheses in a project.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/liam-skillsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skills steer an agent's reasoning and some of them run shell commands to test hypotheses in a project.
This is third-party code. Review the repository files before installing.
What it does
Liam-Skills is a skill pack for Claude Code that breaks decision work into five steps. `osborn` generates many options without judging, `hegel` narrows them through calibrate-and-challenge loops, `bayes` tests conclusions against real code and data, `feynman` hunts for self-deception in the results and `debono` recovers ideas rejected by mistake. The `taixu-pipeline` skill runs all five in sequence and pauses between steps for confirmation. The `--steps hegel,bayes` flag picks only some of them. A separate `socrates` skill examines outside claims for facts, cherry-picked data and hidden assumptions. Skill texts and the README are in Chinese.
Who it is for. Engineers and product managers who ask an agent to assess an architecture or product decision and do not trust the first smooth answer.
Good fit when
- Before picking an architecture, dozens of options are needed instead of the first one that came to mind.
- An agent's assessment sounds convincing and should be checked against real code.
- A report or plan is ready and should be searched for self-deception before release.
- Rejected ideas feel too harshly dismissed and deserve a second look.
- A competitor's analytical report should be broken down into facts and assumptions.
Not a fit when
- The task is small and five stages with pauses would only slow it down.
- The team cannot read Chinese although the skill texts are written in it.
- A quick answer is needed without a long analysis and confirmations between steps.
Example request
Run `taixu-pipeline` on the database migration plan and pause after the `bayes` step for my decision.Limitations
Skill texts and the README are written in Chinese, so a Russian-speaking user should have an agent walk through them beforehand. The repository has no license file although a README badge shows MIT. The last commit is from August 2026. The skills steer an agent's reasoning and do not verify it automatically, so the outcome depends on the model. A full `taixu-pipeline` pass takes a lot of time and tokens, and some skills run shell commands for verification.
How to disable. Delete the `osborn`, `hegel`, `bayes`, `feynman`, `debono`, `socrates` and `taixu-pipeline` folders from `~/.claude/skills/`.
FAQ
Which skills are included?
The five steps `osborn`, `hegel`, `bayes`, `feynman` and `debono`, the `taixu-pipeline` orchestrator and a standalone `socrates` for outside claims. The repository also contains `taixu-debate`.
Can I run only some steps?
Yes. The `--steps hegel,bayes` flag runs only convergence and verification, and each skill also works on its own.
What does the pause between steps do?
The orchestrator shows a status card with elapsed time and deviations. Then you can continue, skip, stop or repeat the step.
Does the pack work with other agents?
The README names Claude Code, Qoder and Aone Copilot. Any platform that follows the `SKILL.md` specification works by copying the skill folders.
Is a shell required?
For some skills, yes: the README asks for an environment that can run Bash because verification uses commands.
Discussion
No questions or reviews yet.
Related
SkillsAbout a hundred skills for Claude Code: meetings, research, publishing, TDD, Telegram, email and personal analytics
The author's personal skills for reading papers and books, dissecting concepts, writing and PNG cards; instructions in Chinese, installed via skills CLI
A mentor skill for Claude Code that guides you to LeetCode solutions with hints instead of handing over code at once
Seven Claude Code skills: turn a video, article or PDF link into text and an action plan, plus a Scrum coach, a session log and an unblock helper
Claude Code Toolkit (robertguss)
Claude Code Toolkit
A set of Claude Code skills and hooks: brainstorming, worker-led project leadership, a wiki, and a book pipeline.
A catalog of 28 mental-model skills for agents: diagnosis, decisions, systems, risk, strategy and idea generation
README in short
Five steps draw on Osborn, Hegel, Bayes, Feynman and de Bono. A separate skill dissects outside claims and an orchestrator joins the steps with pauses.