Crate loopsmith_core 

Config model and validation for loopsmith.

A loop config is four bundles, each answering one question:

• intent — what is this loop for, and how would we know it worked?
• execution — how does the work get done?
• safety — what must not happen, and when does this stop?
• evolution — how is this allowed to change itself?

Until 1.0 these were fourteen flat keys, ten of them known by a letter. Every one of those spellings still loads: parse_str runs the document through config::legacy before typing it, and reports what moved. See config::bundles for why the grouping is what it is.

Validation exists to make the corpus rule enforceable: a goal without a machine-checkable validation is the single most common way loops fail, so the config is rejected rather than run.

Re-exports

pub use md::parse_md;

pub use md::parse_md_reporting;

pub use md::render_md;

pub use validate::validate;

pub use validate::Issue;

pub use validate::Severity;

pub use validate::ValidationReport;

pub use config::*;

Modules

config

The config model, one module per named section.

md

Markdown-native config: the same A–J model, written as a document.

permissions

Permission preflight.

validate

Validation of the four-bundle model.

Enums

CoreError

Functions

is_markdown

Whether a path should be read as a markdown config.

json_schema

The JSON Schema for a loop config, derived from the Rust model.

load

Load a config from Markdown, YAML, or JSON.

load_validated

Load and validate in one step, treating any error-severity issue as fatal.

parse_str

Parse config text, trying YAML first (a superset of JSON in practice) and falling back to strict JSON so both error messages survive to the caller.

parse_str_reporting

parse_str, additionally reporting which 0.3 keys were relocated.