Reasoning framework

`tcx-brain`

Author user-owned inquiry and interpretation principles, then manage immutable installed Brain versions separately.

Manage a user-owned source #

An explicit natural-language request may create or revise investment-brains/<investment-brain-id> and validate it without a lifecycle marker. Authoring uses only user-selected lessons, Wiki pages, and counterexamples, performs privacy review, and abstracts inquiry and interpretation heuristics instead of copying factual or private prose.

Promote selected knowledge

Put the causal questions and falsifiers from this Wiki page into investment-brain-quality-growth.

The revision uses the minimum higher patch version and stops after validation. Deletion and missing privacy choices remain explicit.

Use the managed lifecycle #

List, inspect, and validate are proof-free. Install, update, activate, deactivate, roll back, or remove requires a fresh root task with $tcx-brain on the first meaningful line and uses manage_investment_brain. Source authoring stays ordinary workspace-file work; managed packages and projections are never edited directly.

Install a reviewed source

$tcx-brain
Validate and install investment-brains/investment-brain-quality-growth as inactive.

Installation starts inactive. Activation requires an explicit request. Managed remove drops the Head Manager projection but retains immutable versions for historical provenance; deleting the user-owned source is a separate action.

Select a Brain only for the run that needs it #

Use an active Brain

$investment-brain-quality-growth
Analyze ACME and state the evidence that would falsify the framework.

The exact Brain id may be a plain token or its matching projected workspace skill link. Repeating the same id is harmless; selecting distinct Brain ids fails before analysis.

tcx-brain manages Brains; it does not select one for analysis. A Brain never selects roles, calls tools, changes policy, edits memory, or grants approval, broker, order, or execution authority.