docs: rebuild GUL README as developer landing page

[torakagemusha-sudo]

Summary

Rebuilds the repository README as a focused developer landing page while preserving access to the previous full v2.2 reference through a permanent commit link.

What changed

• leads with the operational problem GUL solves rather than requiring prior formal-methods context;
• positions GUL as an executable four-valued policy and decision layer, not a generic AI CLI;
• adds a verified 60-second installation, validation, and inference path;
• includes the stable output from the existing golden fixture;
• adds practical Python API examples;
• explains the decision algebra, confidence composition, and jurisdiction model;
• documents the Python and native runtime architecture;
• adds executable JSON-spec and native dataset-generation examples;
• introduces an explicit capability-boundary table separating implemented behavior from architectural direction;
• clarifies GUL's intended relationship to TFIR and Helmsdeep without making them package dependencies;
• adds package-map, development, verification, status, and contribution sections;
• links the detailed runtime guide, release specification, project records, curriculum plan, and previous README snapshot.

Why

The previous README was technically comprehensive but front-loaded theory and exposed the practical value too late. This revision reduces evaluation friction for engineers discovering the project while retaining formal precision and truthful implementation-status claims.

Validation

Documentation-only change.

Verified against:

• pyproject.toml for package name, version, Python support, license, and dependency claims;
• .github/workflows/runtime-ci.yml for the Python 3.10–3.13 test matrix and smoke commands;
• cpp/CMakeLists.txt for C++17 and native build details;
• examples/specs/basic_infer.gul.json and tests/golden/basic_infer.infer.json for the quick-start example and stable output;
• docs/runtime_usage.md for runtime boundaries, fallback behavior, fact semantics, supported tags, and dataset constraints.

No runtime code, schemas, tests, or build configuration were changed.

[cursor]

Risk: low. Approved: documentation-only README rewrite with all CI checks passing. Cursor Bugbot was not present on this PR; no reviewers assigned.

  

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: da85bb8b80

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Clarify which missing facts defer

This overstates the atom semantics for fact environments: FactEnvironment.evaluate_predicate only defers missing role/attribute maps and missing custom keys, while absent belongs_to and in_context relations return DENY (facts.py lines 116-128). Users writing membership/context policies from this README could omit a relation expecting escalation, but the runtime will produce an authoritative denial instead.

Useful? React with 👍 / 👎.