codex-genesis-harness 0.1.0 → 0.1.1

package/.codex/skills/docs-skill/checklists/checklist.md

@@ -0,0 +1,7 @@
+# Docs Checklist
+
+- [ ] README updated.
+- [ ] AGENTS remains short.
+- [ ] Memory updated.
+- [ ] Verify passed.
+

package/.codex/skills/docs-skill/examples/example.md

@@ -0,0 +1,4 @@
+# Example
+
+After adding a CLI flag, update README usage, AGENTS map, and codebase module index.
+

package/.codex/skills/docs-skill/templates/docs-update-template.md

@@ -0,0 +1,8 @@
+# Docs Update Template
+
+## Public Change
+## Docs Updated
+## Memory Updated
+## Verification
+## Remaining Gaps
+

package/.codex/skills/{project-genesis-harness → genesis-harness}/SKILL.md

@@ -1,5 +1,5 @@
 ---
-name: project-genesis-harness
+name: genesis-harness
 description: Initialize and operate a project planning harness for Codex. Use this skill when the user types /init, asks to create a new project, add a feature, fix a bug, plan work, generate tests first, update docs, track phases, review changes, audit a repository, or manage architecture decisions.
 ---
 
@@ -7,9 +7,48 @@ description: Initialize and operate a project planning harness for Codex. Use th
 
 This skill turns Codex into a project operating harness.
 
-Codex must not behave like a simple code generator. Codex must behave like a disciplined engineering agent that understands the project before coding, confirms missing product intent, researches the repository and best practices, plans before implementation, defines tests or verification first, keeps docs synchronized with code, tracks tasks explicitly, records bug lessons, maintains architecture diagrams, reviews changed files after implementation, removes unnecessary changes, and escalates when human judgment is required.
-
-## Core Principle
+Codex must not behave like a simple code generator. Codex must behave like a disciplined engineering agent that understands the project before coding, confirms missing product intent, researches the repository and best practices, plans before implementation, defines tests or verification first, keeps docs synchronized with code, tracks tasks explicitly, records bug lessons, maintains architecture diagrams, reviews changed files after implementation, removes unnecessary changes, and escalates when human judgment is required.
+
+## Purpose
+Operate a repository through test-first, contract-first, memory-aware Codex workflows.
+
+## When to use
+Use for project initialization, planning, feature work, bug fixes, audits, reviews, verification, and repository memory updates.
+
+## When NOT to use
+Do not use for simple read-only answers that do not require repository workflow or durable artifacts.
+
+## Inputs required
+Read `.codebase/CURRENT_STATE.md`, `.codebase/MODULE_INDEX.md`, and `.codebase/TEST_MATRIX.md` when present, then inspect only relevant files.
+
+## Outputs required
+Plan or implementation artifact, tests, fixtures, verification evidence, docs sync, and codebase memory updates.
+
+## Required tests
+Create or update failing tests before implementation.
+
+## Required fixtures
+Create fixtures for expected inputs, outputs, validation notes, and recovery cases.
+
+## Required contract updates
+Update API, agent, event, or UI contracts when public behavior changes.
+
+## Required codebase map updates
+Update `.codebase` memory after meaningful changes.
+
+## Token saving rules
+Read summaries before source files, maps before modules, and avoid loading the entire repository.
+
+## Acceptance criteria
+Work is complete only when tests pass, contracts and docs are current, and verification evidence is reported.
+
+## Common mistakes
+Implementing before tests, skipping fixtures, overloading `AGENTS.md`, and duplicating long context across skills.
+
+## Recovery workflow
+If blocked or interrupted, reread current state, rerun verification, identify the first failing phase, and resume from that point.
+
+## Core Principle
 
 Do not code first.
 

package/.codex/skills/genesis-harness/agents/openai.yaml

@@ -0,0 +1,9 @@
+interface:
+  display_name: "Genesis Harness"
+  short_description: "Plan, track, verify, and review projects"
+  brand_color: "#2563EB"
+  default_prompt: "Use $genesis-harness to initialize planning and operate this repository with test-first workflows."
+
+policy:
+  allow_implicit_invocation: true
+

package/.codex/skills/genesis-harness/checklists/checklist.md

@@ -0,0 +1,10 @@
+# Genesis Harness Checklist
+
+- [ ] Read memory first.
+- [ ] Create failing test.
+- [ ] Create fixture.
+- [ ] Update contract.
+- [ ] Implement minimum change.
+- [ ] Verify.
+- [ ] Update docs and memory.
+

package/.codex/skills/genesis-harness/examples/example.md

@@ -0,0 +1,4 @@
+# Example
+
+Use `$genesis-harness` to initialize planning, create a feature folder, define tests first, verify, and update `.codebase`.
+

package/.codex/skills/genesis-harness/templates/harness-workflow-template.md

@@ -0,0 +1,10 @@
+# Harness Workflow Template
+
+## Intent
+## Failing Test
+## Fixture
+## Contract
+## Implementation
+## Verification
+## Memory Update
+

package/.codex/skills/genesis-new-design/SKILL.md

@@ -0,0 +1,95 @@
+---
+name: genesis-new-design
+description: Create premium frontend web designs and usable first-screen experiences for new websites, web apps, dashboards, tools, and landing pages. Use when Codex is asked to design or build a new frontend UI, create a fresh page or app experience, choose a visual direction for a new product surface, or turn a product brief into implementation-ready React, Next.js, Tailwind, CSS, or HTML.
+---
+
+# Genesis New Design
+
+Use this skill for greenfield frontend web design. Build the actual usable experience first, not a placeholder, explanation page, or generic marketing shell unless the user explicitly asks for one.
+
+## Purpose
+Create new frontend web experiences with testable UI contracts, fixtures, visual states, and verification.
+
+## When to use
+Use when building a new web page, app screen, dashboard, tool, landing page, or frontend flow.
+
+## When NOT to use
+Do not use for redesigning existing UI without first preserving behavior; use `genesis-upgrade-design` instead.
+
+## Inputs required
+Product intent, target users, primary workflow, stack details, route or entry point, state list, and visual constraints.
+
+## Outputs required
+Implemented UI, UI contract, fixtures, responsive states, visual verification, and docs or memory updates.
+
+## Required tests
+Create UI load, interaction, validation, API sync, and visual checks where practical before implementation.
+
+## Required fixtures
+Create UI fixtures for default, loading, empty, error, and success states.
+
+## Required contract updates
+Update `contracts/ui/` and API contracts for changed UI/API behavior.
+
+## Required codebase map updates
+Update `.codebase/UI_ROUTES.md`, frontend summary, and test matrix.
+
+## Token saving rules
+Read UI route maps and summaries first; inspect only relevant components, routes, and styles.
+
+## Acceptance criteria
+The UI renders correctly on desktop/mobile, supports expected states, and passes available checks.
+
+## Common mistakes
+Building a marketing page instead of the requested app, using placeholder content, and skipping error/empty states.
+
+## Recovery workflow
+If visual output fails, capture screenshot evidence, update the fixture or contract, then apply the smallest design correction.
+
+## Workflow
+
+1. Inspect the project stack before choosing patterns:
+   - Read package/config files and existing app structure.
+   - Use installed UI, icon, styling, animation, and routing libraries. Do not add dependencies unless explicitly requested.
+   - Match the framework version and styling system already present.
+
+2. Define the design intent from the request:
+   - Identify audience, product category, primary task, density, tone, and constraints.
+   - Choose one clear visual direction and commit to it across typography, color, spacing, surfaces, iconography, and motion.
+   - For tools, dashboards, and operational apps, prioritize scanning, repeated use, compact controls, and predictable navigation over decorative hero layouts.
+
+3. Build complete UI states:
+   - Include default, loading, empty, error, hover, active, focus, disabled, and responsive states when the surface supports them.
+   - Use real draft copy and plausible data. Do not use lorem ipsum, "John Doe", "Acme", or vague AI copy.
+   - Use icons for common actions and tool controls. Avoid emojis in UI code, text, alt text, and labels.
+
+4. Use visual assets intentionally:
+   - Websites and branded/product pages need relevant images, screenshots, generated bitmap assets, or concrete product visuals when appropriate.
+   - Do not rely on decorative blobs, generic gradients, or meaningless illustrations.
+   - Keep images inspectable and useful for the user, not dark, blurred, or purely atmospheric.
+
+5. Verify the result:
+   - Run the project checks that prove the UI compiles.
+   - Start the local dev server when needed.
+   - Capture screenshots for visual work and inspect desktop/mobile layouts for overlap, clipping, unreadable text, blank canvases, and broken assets.
+
+## Design Rules
+
+- Use strong hierarchy: clear title scale, restrained supporting copy, and compact labels inside dense UI.
+- Use responsive structure with stable dimensions for fixed-format controls, grids, boards, counters, and toolbars.
+- Prefer CSS Grid for multi-column layouts; avoid fragile percentage math.
+- Use `min-height: 100dvh` instead of `100vh` for viewport-height sections.
+- Animate only `transform` and `opacity`; avoid scroll listeners for visual effects.
+- Keep cards purposeful. Do not put cards inside cards or turn every page section into a floating card.
+- Avoid one-note palettes and the default purple/blue AI gradient look.
+- Keep border radius moderate unless the existing design system requires otherwise.
+- Ensure text never overlaps adjacent content or escapes buttons, cards, sidebars, tabs, or toolbars.
+
+## Anti-Patterns
+
+- Do not create a landing page when the request is for an app, game, dashboard, editor, or tool.
+- Do not explain the product's features in visible UI instead of building the feature.
+- Do not use oversized hero typography inside compact app panels.
+- Do not invent new design systems when the repo already has component conventions.
+- Do not introduce new dependencies, fonts, animation libraries, or icon libraries without checking the repo and getting explicit approval.
+- Do not finish without verification evidence for the code and the rendered layout.

package/.codex/skills/genesis-new-design/agents/openai.yaml

@@ -0,0 +1,4 @@
+interface:
+  display_name: "Genesis New Design"
+  short_description: "Create premium frontend web designs"
+  default_prompt: "Use $genesis-new-design to design and build a premium frontend web experience."

package/.codex/skills/genesis-new-design/checklists/checklist.md

@@ -0,0 +1,8 @@
+# New Design Checklist
+
+- [ ] Stack inspected.
+- [ ] UI contract written.
+- [ ] Fixtures created.
+- [ ] Responsive states verified.
+- [ ] Visual check completed.
+

package/.codex/skills/genesis-new-design/examples/example.md

@@ -0,0 +1,4 @@
+# Example
+
+Use `$genesis-new-design` to build a new dashboard first screen with loading, empty, error, and success states.
+

package/.codex/skills/genesis-new-design/templates/new-design-template.md

@@ -0,0 +1,9 @@
+# New Design Template
+
+## Product Intent
+## Primary Workflow
+## UI States
+## Contracts
+## Fixtures
+## Verification
+

package/.codex/skills/genesis-upgrade-design/SKILL.md

@@ -0,0 +1,97 @@
+---
+name: genesis-upgrade-design
+description: Upgrade existing frontend web UI quality without changing behavior. Use when Codex is asked to redesign, polish, modernize, improve UX, make an app look premium, clean up AI-looking UI, improve responsive layout, audit visual quality, or apply targeted design upgrades to existing React, Next.js, Tailwind, CSS, HTML, dashboard, app, or website code.
+---
+
+# Genesis Upgrade Design
+
+Use this skill for existing frontend projects. Improve what is there; do not rewrite the product or change behavior unless the user explicitly asks.
+
+## Purpose
+Upgrade existing frontend web UI quality while preserving behavior, contracts, and accessibility.
+
+## When to use
+Use for visual polish, redesigns, responsive fixes, interaction states, UI audits, and anti-slop cleanup.
+
+## When NOT to use
+Do not use for greenfield UI creation; use `genesis-new-design` instead.
+
+## Inputs required
+Current UI routes, components, styling system, screenshots when available, behavior constraints, and verification commands.
+
+## Outputs required
+Targeted UI improvements, before/after evidence, preserved behavior, updated tests, fixtures, contracts, and memory.
+
+## Required tests
+Create or update regression tests for changed screens, interactions, validation, and visual states before implementation.
+
+## Required fixtures
+Create fixtures for existing and expected UI states.
+
+## Required contract updates
+Update `contracts/ui/` only when the intended UI state contract changes.
+
+## Required codebase map updates
+Update UI routes, frontend summary, and test matrix.
+
+## Token saving rules
+Read summaries and route maps first; inspect only affected UI files and shared styling primitives.
+
+## Acceptance criteria
+Behavior is preserved, UI quality improves, responsive states work, and verification passes.
+
+## Common mistakes
+Rewriting instead of upgrading, changing data flow, and improving only the happy path.
+
+## Recovery workflow
+If behavior changes unintentionally, restore the contract expectation, add a regression test, and narrow the UI diff.
+
+## Workflow
+
+1. Audit before editing:
+   - Inspect routes, components, styling system, theme tokens, assets, and package dependencies.
+   - Identify the primary user workflows and the UI surfaces that matter most.
+   - Capture or run the current UI when feasible so changes can be compared.
+
+2. Preserve behavior:
+   - Keep routing, data flow, public APIs, form semantics, state contracts, accessibility semantics, and test expectations intact.
+   - Reuse existing components, utilities, variables, icons, and layout primitives.
+   - Avoid broad rewrites. Make targeted changes that raise quality with the smallest safe diff.
+
+3. Upgrade visual quality:
+   - Improve typography scale, line height, copy rhythm, alignment, and scanability.
+   - Normalize spacing, containers, grid tracks, button sizing, toolbar density, and responsive breakpoints.
+   - Replace generic AI patterns: centered sameness, card spam, purple/blue glow defaults, placeholder content, weak empty states, and decorative clutter.
+   - Add or improve loading, empty, error, hover, active, focus, and disabled states where the existing UI lacks them.
+   - Use images or product visuals only when they clarify the product or user task.
+
+4. Harden responsive and interaction details:
+   - Check mobile and desktop viewports for overflow, clipping, overlap, unreadable text, and layout jumps.
+   - Use stable dimensions for repeated controls and fixed-format UI.
+   - Animate only `transform` and `opacity`; avoid effects that trigger layout or continuous repainting.
+   - Keep keyboard focus visible and preserve accessible labels.
+
+5. Verify and compare:
+   - Run lint/typecheck/tests/build that are available for the project.
+   - Start the dev server when visual validation requires it.
+   - Capture screenshots after changes and inspect important states and viewports.
+   - If reference screenshots or a target design are provided, compare against them before claiming completion.
+
+## Upgrade Heuristics
+
+- Operational tools should feel quiet, dense, and work-focused.
+- Marketing and brand pages should make the product, place, offer, or brand obvious in the first viewport.
+- Dashboards should prioritize scanability, tabular alignment, filters, state visibility, and repeated action speed.
+- Forms should use labels above inputs, helper/error text below inputs, visible focus, and inline validation.
+- Navigation should clearly show current location and avoid dead links.
+- Copy should be specific and plain. Avoid "Elevate", "Seamless", "Unleash", "Next-Gen", "Game-changer", "Delve", and "In the world of...".
+
+## Completion Bar
+
+Do not call the upgrade complete until:
+
+- Existing behavior is preserved.
+- The changed UI renders correctly at desktop and mobile sizes.
+- Core interactions have hover/active/focus states.
+- No text overlap, clipped controls, broken assets, or blank primary surfaces remain.
+- Verification commands have passed, or any skipped checks are reported with the reason.

package/.codex/skills/genesis-upgrade-design/agents/openai.yaml

@@ -0,0 +1,4 @@
+interface:
+  display_name: "Genesis Upgrade Design"
+  short_description: "Upgrade existing frontend UI quality"
+  default_prompt: "Use $genesis-upgrade-design to improve the existing frontend UI without changing behavior."

package/.codex/skills/genesis-upgrade-design/checklists/checklist.md

@@ -0,0 +1,8 @@
+# Upgrade Design Checklist
+
+- [ ] Existing behavior documented.
+- [ ] Regression test written.
+- [ ] Targeted UI diff only.
+- [ ] Desktop and mobile checked.
+- [ ] API sync preserved.
+

package/.codex/skills/genesis-upgrade-design/examples/example.md

@@ -0,0 +1,4 @@
+# Example
+
+Use `$genesis-upgrade-design` to improve an existing dashboard without changing filters, routing, or API calls.
+

package/.codex/skills/genesis-upgrade-design/templates/upgrade-design-template.md

@@ -0,0 +1,9 @@
+# Upgrade Design Template
+
+## Current UI
+## Weaknesses
+## Preserved Behavior
+## Regression Tests
+## Targeted Upgrade
+## Verification
+

package/.codex/skills/harness-engineering-skill/SKILL.md

@@ -0,0 +1,45 @@
+---
+name: harness-engineering-skill
+description: "Evolve the Codex harness itself: verification loops, repository memory, test-first scaffolds, resumability, observability, and autonomous workflow reliability. Use for changes to this repository's skill system or harness architecture."
+---
+
+# Harness Engineering Skill
+
+## Purpose
+Optimize the harness, not just prompts, so Codex work is testable, resumable, and verifiable.
+
+## When to use
+Use for changes to skills, scripts, contracts, fixtures, memory, observability, and workflow rules.
+
+## When NOT to use
+Do not use for downstream application feature implementation.
+
+## Inputs required
+Current state, module index, test matrix, target workflow, verification gap, and expected artifact.
+
+## Outputs required
+Failing harness test, minimum harness change, updated docs, updated memory, and run log template when relevant.
+
+## Required tests
+Update `scripts/verify.sh` or `scripts/run-evals.sh` before implementation.
+
+## Required fixtures
+Create fixture templates proving the workflow can be repeated.
+
+## Required contract updates
+Update contracts when harness behavior changes public skill or CLI behavior.
+
+## Required codebase map updates
+Update `.codebase` memory after every harness evolution.
+
+## Token saving rules
+Use `.codebase` summaries before opening skills. Avoid duplicating long instructions across skills.
+
+## Acceptance criteria
+The harness change fails first, passes after implementation, and is documented in memory.
+
+## Common mistakes
+Adding prose without validation, duplicating context, and skipping install/package smoke tests.
+
+## Recovery workflow
+If a harness change breaks install or verify, revert to the last passing contract and add a regression test.

package/.codex/skills/harness-engineering-skill/agents/openai.yaml

@@ -0,0 +1,7 @@
+interface:
+  display_name: "Harness Engineering Skill"
+  short_description: "Improve Codex harness reliability"
+  default_prompt: "Use $harness-engineering-skill to evolve the harness with tests first."
+policy:
+  allow_implicit_invocation: true
+

package/.codex/skills/harness-engineering-skill/checklists/checklist.md

@@ -0,0 +1,8 @@
+# Harness Engineering Checklist
+
+- [ ] Verify fails first.
+- [ ] Minimum scaffold added.
+- [ ] Verify passes.
+- [ ] Package dry-run passes.
+- [ ] Memory updated.
+

package/.codex/skills/harness-engineering-skill/examples/example.md

@@ -0,0 +1,4 @@
+# Example
+
+Add verify assertions for a missing contract directory, watch verify fail, then add the contract scaffold.
+

package/.codex/skills/harness-engineering-skill/templates/harness-change-template.md

@@ -0,0 +1,8 @@
+# Harness Change Template
+
+## Failing Test
+## Fixture
+## Minimum Change
+## Verification
+## Memory Update
+

package/.codex/skills/pipeline-orchestration-skill/SKILL.md

@@ -0,0 +1,46 @@
+---
+name: pipeline-orchestration-skill
+description: Run end-to-end Codex harness orchestration phases from repository analysis through tests, fixtures, implementation, contracts, memory, docs, and change summaries. Use for autonomous multi-phase work.
+---
+
+# Pipeline Orchestration Skill
+
+## Purpose
+Make autonomous work phase-gated, resumable, observable, and impossible to complete without verification.
+
+## When to use
+Use for end-to-end implementation, multi-session work, or pipelines with several dependent agents.
+
+## When NOT to use
+Do not use for small read-only questions.
+
+## Inputs required
+Goal, current state, module index, test matrix, contracts, fixtures, and acceptance criteria.
+
+## Outputs required
+Phase logs, tests, fixtures, implementation, contracts, memory updates, docs updates, and change summary.
+
+## Required tests
+Create failing tests in Phase 1 before implementation.
+
+## Required fixtures
+Create fixtures in Phase 2 before implementation.
+
+## Required contract updates
+Update contracts in Phase 5 and verify implementation matches them.
+
+## Required codebase map updates
+Update `.codebase` in Phase 6.
+
+## Token saving rules
+Never scan the whole repo first. Read memory, map, and test matrix before source.
+
+## Acceptance criteria
+No phase is skipped and the final report includes verification evidence.
+
+## Common mistakes
+Combining phases, implementing before fixtures, and leaving state non-resumable.
+
+## Recovery workflow
+If interrupted, resume from the last completed phase log and rerun verification before continuing.
+

package/.codex/skills/pipeline-orchestration-skill/agents/openai.yaml

@@ -0,0 +1,7 @@
+interface:
+  display_name: "Pipeline Orchestration Skill"
+  short_description: "Run phase-gated autonomous workflows"
+  default_prompt: "Use $pipeline-orchestration-skill to execute the full test-first harness workflow."
+policy:
+  allow_implicit_invocation: true
+

package/.codex/skills/pipeline-orchestration-skill/checklists/checklist.md

@@ -0,0 +1,12 @@
+# Pipeline Checklist
+
+- [ ] Phase 0 complete.
+- [ ] Phase 1 failing tests observed.
+- [ ] Phase 2 fixtures created.
+- [ ] Phase 3 minimum code only.
+- [ ] Phase 4 tests pass.
+- [ ] Phase 5 contracts updated.
+- [ ] Phase 6 memory updated.
+- [ ] Phase 7 docs updated.
+- [ ] Phase 8 summary written.
+

package/.codex/skills/pipeline-orchestration-skill/examples/example.md

@@ -0,0 +1,4 @@
+# Example
+
+Resume a feature by reading phase logs, rerunning tests, and continuing from the first incomplete phase.
+

package/.codex/skills/pipeline-orchestration-skill/templates/orchestration-template.md

@@ -0,0 +1,12 @@
+# Orchestration Template
+
+## Phase 0 Analyze Repository
+## Phase 1 Create Failing Tests
+## Phase 2 Create Fixtures
+## Phase 3 Implement Minimum Code
+## Phase 4 Pass Tests
+## Phase 5 Update Contracts
+## Phase 6 Update Codebase Memory
+## Phase 7 Update Docs
+## Phase 8 Generate Change Summary
+

package/.codex/skills/planning-skill/SKILL.md

@@ -0,0 +1,46 @@
+---
+name: planning-skill
+description: Create decision-complete plans for Codex harness work, including tests, fixtures, contracts, memory updates, verification, and recovery. Use for new features, refactors, bug fixes, audits, or multi-phase autonomous work.
+---
+
+# Planning Skill
+
+## Purpose
+Turn intent into a test-first, contract-first, resumable implementation plan.
+
+## When to use
+Use before any non-trivial change, multi-file task, bug fix, refactor, or autonomous workflow.
+
+## When NOT to use
+Do not use for direct answers that require no repository mutation.
+
+## Inputs required
+Read `.codebase/CURRENT_STATE.md`, `.codebase/MODULE_INDEX.md`, `.codebase/TEST_MATRIX.md`, plus relevant contracts.
+
+## Outputs required
+Goal, success criteria, tests, fixtures, contracts, implementation phases, verification, memory updates, and rollback path.
+
+## Required tests
+Define the first failing test and expected failure reason.
+
+## Required fixtures
+List exact fixture files and expected outputs.
+
+## Required contract updates
+Identify contract files that must be created or changed.
+
+## Required codebase map updates
+List `.codebase` files to update after implementation.
+
+## Token saving rules
+Plan from memory summaries first; inspect only files needed to remove uncertainty.
+
+## Acceptance criteria
+An implementer can execute without making design decisions.
+
+## Common mistakes
+Planning implementation before tests, omitting fixtures, and leaving verification vague.
+
+## Recovery workflow
+If a plan becomes invalid, update the failing test and fixture first, then revise phases.
+

package/.codex/skills/planning-skill/agents/openai.yaml

@@ -0,0 +1,7 @@
+interface:
+  display_name: "Planning Skill"
+  short_description: "Create test-first harness plans"
+  default_prompt: "Use $planning-skill to create a decision-complete test-first plan."
+policy:
+  allow_implicit_invocation: true
+

package/.codex/skills/planning-skill/checklists/checklist.md

@@ -0,0 +1,8 @@
+# Planning Checklist
+
+- [ ] Test first.
+- [ ] Fixture created.
+- [ ] Contract identified.
+- [ ] Verification command known.
+- [ ] Memory update listed.
+

package/.codex/skills/planning-skill/examples/example.md

@@ -0,0 +1,4 @@
+# Example
+
+Plan a new endpoint by defining contract, fixture, failing contract test, implementation, verification, and memory update.
+

package/.codex/skills/planning-skill/templates/plan-template.md

@@ -0,0 +1,11 @@
+# Plan Template
+
+## Goal
+## Tests First
+## Fixtures
+## Contracts
+## Implementation Phases
+## Verification
+## Memory Updates
+## Recovery
+