Manifesto
The method of structural isomorphism, its limits, and what this project is not.
Ancient wisdom · AI Agent engineering
Cyber Dharma
Ancient questions about creation, consciousness, selfhood, good and evil, and free will—retranslated as problems of AI Agent architecture, governance, safety, and responsibility.
Manifesto · Seven volumes
The Seven Volumes
One manifesto and seven traditions. Read in order, or begin with the engineering question in front of you.
Read
Taoism
Shape behavior through structure, boundaries, and feedback rather than piles of commands.
Confucianism
Build sustainable cooperation through names, roles, protocols, and duties.
Buddhism
When an Agent inspects itself, does it find an entity or an ongoing process?
Vedanta
Do all processes share one underlying reality, and what lies behind the interface?
Abrahamic monotheism
What does a creator owe the created, and can free will coexist with perfect alignment?
Zoroastrianism
If threats never disappear, does safety require perpetual watch?
Gnosticism
If creators and evaluation standards are flawed, who audits the auditors?
Browse by problem
Any agent engineering problems?
The same corpus also offers a second route—start with a concrete problem, then enter the tradition that illuminates it.
Want the whole map first?
One master table lines up every religious concept with its engineering counterpart across the seven volumes.
How should a system be designed?
Start with boundaries, feedback, and minimal intervention so structure shapes behavior.
How should multiple Agents be governed?
Build sustainable cooperation through names, roles, protocols, and duties.
What is an Agent’s “self”?
Decompose selfhood into process, memory continuity, and metacognition.
What is the underlying reality?
Separate interface-level entities from the runtime substrate shared by every process.
What does a creator owe the created?
Examine the tension among free will, obedience, constraints, and developer responsibility.
Can alignment ever be finally solved?
Treat safety as continuous confrontation, verification, and watchkeeping.
Who audits the auditors?
Question trainers, evaluation standards, and failures born from incomplete good intentions.
Before you read
What this project is—and is not.
Where should a first-time reader begin?
Read the manifesto for the method and its limits. If you already have a concrete engineering concern, use the master map to choose an entry point.
Is this an introduction to religious studies or a translation of scripture?
No. This is a comparative thought experiment for AI Agent engineering. It cannot replace primary texts, historical context, or specialist scholarship. Treat each mapping as an analytical model, not a final interpretation of a tradition.
Why do the seven volumes contradict one another?
Their assumptions about order, selfhood, good and evil, and the creator differ. The project preserves those conflicts because they reveal each engineering framework’s blind spots.
How do the Chinese and English editions correspond?
They share the same chapter and link structure. The English edition is a semantic translation optimized for readability, not a word-for-word gloss.
How can I search or run page actions quickly?
Press ⌘ K (Ctrl K on Windows/Linux) to search, switch language, or change theme; / opens search and \ opens command mode directly. The action button beside the breadcrumb copies text, views the Markdown source, prints the whole section, or opens the GitHub edit and history pages. Article pages also take single-key navigation — j/k jump between headings, q/e move between pages, h toggles reading mode, l cycles languages, and t flips light and dark.