Skip to content

Releases: integrii/logic-patterns

v1.0.5

Choose a tag to compare

@integrii integrii released this 13 Oct 09:31

Release the latest local skill versions with refreshed documentation and discovery metadata.

  • Verify all 18 skills match their latest installed Codex copies and pass skill validation.
  • Refresh the root and bundle READMEs to describe the current workflows.
  • Add five-whys and github-release-builds to the discovery index and correct the skill count to 18.
  • Set both plugin bundles and the root manifest to version 1.0.5.

Updates since v1.0.4 include bounded implementation planning without shipping procedures, complete-batch adversary reviews, GPT-6 Astra medium reviewers for plan execution, and workspace-root PersonaStack plan storage.

v1.0.4

Choose a tag to compare

@integrii integrii released this 02 Sep 20:35

Require implement-plan to run a clean adversary loop after code implementation and focused tests pass before closing a phase or plan.

v1.0.3

Choose a tag to compare

@integrii integrii released this 04 Aug 22:44

Patch notes

  • Validate LAN UI candidates inside a session-scoped Golden Stack through port-forwarding with an ephemeral account.
  • Keep Golden Stack findings batched before implementation fixes.
  • Assign UX quality gates to a 5.6 Sol medium reviewer, 5.6 Terra planner, and 5.6 Luna implementer.
  • Select 5.6 Terra or 5.6 Sol at medium or high strength for plan creation based on sensitivity and complexity.

v1.0.2

Choose a tag to compare

@integrii integrii released this 31 May 10:00

Changelog

  • Relaxed session-review to capture durable patterns whenever they exist, without a minimum required count.
  • No behavior changes; docs-only release.

Add session-review skill

Choose a tag to compare

@integrii integrii released this 31 May 09:58

Changelog

  • Added new session-review skill in plugins/logic-patterns/skills/session-review/SKILL.md.
  • Captures trigger, macro patterns, and reusable checklists for converting session discoveries into repo-specific skills.
  • Added discoverability entries in plugins/logic-patterns/README.md and the top-level README.md plugin list.
  • No code behavior changes; docs-only release.

v1.0.0

Choose a tag to compare

@integrii integrii released this 31 May 09:24

Added

  • Initial public plugin release for logic-patterns and development-patterns.
  • Logic plugin skills: 1-3-1, adversary-loop, bifurcate, first-principles-rebuild, gaslight-loop, greyhat, make-a-plan, multidimensional-planning.
  • Development plugin skills: architecture-decision-records, architecture-spec, centralized-fix-selection, ephemeral-testing, implement-plan, plan-checklists, spec-driven-development.
  • Plugin manifests with proper .codex-plugin/plugin.json metadata and logos.
  • Root README with install guidance and plugin/skill catalog.

Changed

  • Reorganized repository into two plugins: logic-patterns and development-patterns.
  • Standardized skill documentation and references to local plugin paths.
  • Updated distribution model to support direct Codex marketplace installation.

Removed

  • Removed outdated or duplicate/deprecated skills from this plugin package (go-coding-practices, ts-coding-practices, skill-routing-orchestrator, and other non-local references).
  • Removed hardcoded/incompatible external references and user-specific paths from distribution content.

Operations

  • Added explicit installation path that does not require skill-by-skill configuration.
  • Simplified install workflow to two concise CLI options for plugin users.