mcp_philosophy 0.1.2

Value principle, judgment direction, and ethos layer for the MakeMind Knowledge System providing philosophy-based pipeline intervention and evaluation.

0.1.2 - 2026-06-23 - deterministic prohibition enforcement (forbiddenPatterns)

Changed

• _detectViolation now enforces Prohibition.forbiddenPatterns (mcp_bundle 0.4.4) first — any declared forbidden pattern present (case-insensitive) in the proposed output/action is a deterministic violation. Previously the structural evaluator recognized only two hardcoded NL statement shapes ('uncertain'+'certain', 'hide'+'limitation') and silently returned "not violated" for every other prohibition — so a real prohibition like "never output X" was never caught by checkProhibitions / the post-generation intervene gate (it fell open). forbiddenPatterns is the sound, LLM-free enforcement hook: an author / LLM declares the concrete strings that must never appear and the engine blocks them deterministically. The two built-in heuristics remain as a best-effort fallback; semantic judgment of arbitrary NL statements is explicitly deferred to an LLM seam (spec platform/12-flowbrain-runtime.md §3). Test: test/evaluation/philosophy_evaluator_test.dart (forbiddenPatterns hard/soft + no-match).

Changed (dependency floor)

• mcp_bundle ^0.4.0 → ^0.4.4 — uses Prohibition.forbiddenPatterns, introduced in 0.4.4; floored to guarantee it.

0.1.1 - 2026-05-23 - mcp_bundle 0.4.0 cascade

Changed (cascade)

• mcp_bundle caret bumped from ^0.3.0 to ^0.4.0. mcp_philosophy does not touch UiSection.pages directly, so this release is a caret-only cascade. Consumers should bump to ^0.1.1.

0.1.0 - 2026-04-28 - Initial Release

Added

• PhilosophyEngine — sole implementation of mcp_bundle's PhilosophyPort.
• Philosophy evaluation, prohibition checking, pipeline intervention (pre / during / post stages), tension detection, and ethos evolution proposals.
• Dynamic state weighting integration.
• Re-exports of contract types from mcp_bundle.