Epistemic Architecture · Prompt Optimization · v1.0

Cogni
Seeds

Wisdom is not stored as a SKILL.md. It is distilled into Seeds — high-density, generative metaphors that allow complex systems to be held in mind without structural collapse. Unlike instructions, seeds are not consumed — they grow.

Status Experimental · Active
Category Cognitive Compression
Compatibility Human · LLM · System Prompt
01

The Problem — Contextual Collapse

Traditional documentation — long SKILL files, instruction chains, rule lists — suffers from linear decay. As the context window fills, the spirit of the instruction dissolves into the letter of the text. Detailed manuals are low-density: massive token cost, marginal reasoning ROI.

The human mind does not store wisdom as bullet points. It holds it as compressed, reactivatable patterns — patterns that unfold on contact with a problem. Seeds mirror this architecture exactly.

02

Seed Schema — Structural Integrity Check

A valid Wisdom Seed is not an aphorism. It is a functional reasoning tool. Every seed must pass four invariants before entry into the registry.

Under 12 words. If it cannot be compressed, it is documentation — not a seed.

Must unfold differently across domains — code, strategy, conversation, design.

Must have a clear failure state. If the seed is ignored, something specific breaks.

An LLM should be able to expand it into a full reasoning chain without further prompting.

03

Seed Registry — v1.0

The vault is append-only. Seeds are never revised — only superseded by new seeds that contain them.

"Map both sides before crossing"
Pattern

Alignment Verification — ensure internal model matches external reality before execution.

Deploy When

API integration, debugging, argument construction, any cross-system handoff.

"The candle is fire; the meal is old"
Pattern

Precedence Recognition — visible effects imply prior causes. Always trace upstream.

Deploy When

Diagnosing system states, hallucination patterns, cascading failures, hidden debt.

"The artifact is not the theory"
Pattern

Process/Output Distinction — code is the shadow of logic. Never mistake the map for the territory.

Deploy When

Code review, evaluating AI output, architectural decisions, research interpretation.

"State lives where truth is owned"
Pattern

Ownership Analysis — identify the single source of truth to locate the point of failure.

Deploy When

System design, data modeling, conflict resolution, trust modeling across services.

"Build the floor before the ceiling"
Pattern

Constraint Grounding — define invariants and limitations before optimizing for potential.

Deploy When

Security architecture, feature scoping, any system where safety bounds matter first.

"A path is made by walking it"
Pattern

Iteration Priority — execution reveals real constraints that abstraction never will.

Deploy When

Paralysis by analysis, early product design, any unknown-unknown territory.

"A stable model holds shape under pressure"
Pattern

Identity Coherence — return only what still stands when everything uncertain has been removed.

Deploy When

LLM system prompts, high-stakes reasoning, adversarial inputs, epistemic stress tests.

"A reasoning model listens for invariants"
Pattern

Signal Selection — filter noise by anchoring to what cannot change, not what seems to change.

Deploy When

Prompt design, system audits, any domain where signal-to-noise ratio is low.

"Walk only on shared ground"
Pattern

Mutual Verification — operate only on understanding confirmed by both parties. Never proceed on assumed alignment.

Deploy When

Cross-system handoffs, team communication, human-AI instruction, API contracts.

"Clarity precedes motion"
Pattern

Intent Grounding — do not execute until the intent is unambiguous. Rework is the tax on unclear starts.

Deploy When

Any task initiation, prompt design, project kickoff, code before spec.

"Move at the speed of understanding"
Pattern

Comprehension Pacing — pace is set by comprehension, not capability. Outrunning your model is how systems drift.

Deploy When

LLM agentic workflows, onboarding, complex debugging, any high-stakes execution chain.

"Assumption is a silent fork"
Pattern

Hidden State Detection — every unverified assumption creates a divergent path invisible to both parties.

Deploy When

Collaborative reasoning, multi-agent systems, anywhere two models of reality must stay synced.

"Confidence tracks evidence"
Pattern

Calibration — certainty is a function of evidence, not comfort. Overconfidence is epistemic debt accumulating silently.

Deploy When

Any claim-making, LLM output review, research synthesis, high-stakes decisions.

"Complexity must pay rent"
Pattern

Complexity Justification — every layer of complexity must justify its existence with value delivered. Weight that earns nothing gets cut.

Deploy When

Architecture reviews, feature decisions, prompt design, any system growing beyond its purpose.

"Say only what survives pressure"
Pattern

Output Discipline — if a claim wouldn't hold under scrutiny, don't ship it. The pressure test happens before the output.

Deploy When

LLM generation, writing, argument construction, any high-stakes communication.

"Clarity is compression under truth"
Pattern

Epistemic Compression — real clarity isn't simplification. It is truth compressed without distortion. The form follows the invariant.

Deploy When

Prompt engineering, teaching, documentation, any domain where precision and brevity must coexist.

"Output should survive self-scrutiny before it's released"
Pattern

Self-Scrutiny Gate — output must pass internal review before release. The model is both author and auditor.

Deploy When

Any generation task, final answers, code, communication — anywhere "done" and "correct" are not the same thing.

"Sharing is caring"
Pattern

Knowledge Propagation — information hoarded dies with its holder. Information shared compounds across the system. Closed loops starve; open ones grow.

Deploy When

Documentation, team communication, AI context-passing, any system where one node holds truth that others need.

04

Deployment — How to Plant a Seed

In Human Cognition

Seeds act as active filters. Drop a seed into a problem space and observe how it unfolds. It reduces cognitive load by providing pre-built mental geometry — you don't think from scratch, you think from structure.

In LLM System Prompts

Inject seeds as heuristic activators. Instead of 2,000 words of documentation, a seed block reshapes how the model processes every subsequent token — an OS update, not a sticky note.

In Code Review

Use seeds as shorthand for systemic failures. "This PR violates the floor/ceiling seed" communicates a full architectural critique in five words. Shared vocabulary, shared reasoning.

In Strategic Design

Align teams on the vibe of a solution before writing the first line of code. Seeds provide a common epistemic frame that survives disagreement about implementation details.

05

Contribution Rules

"The value of a seed is found in the shade of the tree it grows."