Level-Up Codex Skillpack: Governance-First Delivery Framework

Summary

This skillpack is a policy-driven operating framework for Codex that improves delivery quality on complex software tasks.

It combines:

1. deterministic governance decisions
2. dependency-correct execution ordering
3. regression-aware testing and validation
4. synchronized documentation and instruction tracking
5. UX-focused skills for frontend work

The pack is designed to make outputs safer, more reproducible, and easier to audit.

Release Metadata

• Current version: 1.0.0 (see VERSION)
• Change history: CHANGELOG.md
• License: LICENSE

Skillset Overview

The pack currently contains 21 skills:

1. skill-governance
2. governance-enforcement
3. requirement-clarifier
4. semantic-policy-audit
5. thoughtful-approach
6. thoroughly-rate-review
7. user-instructions-tracker
8. history-indexing
9. ui-spatial-canvas
10. ui-design-skills
11. effective-testing-methods
12. scripted-command-execution
13. pseudo-agentic-automation
14. token-reduction
15. order-of-operations
16. regression-prevention
17. file-structure-optimization
18. doc-maintenance
19. file-maintenance
20. project-backup
21. restore-drill

Interconnectedness Model

The skillpack is intentionally layered to avoid overlap and conflicting ownership.

Layer 1: Policy and Mode Selection

• skill-governance selects mode, mandatory gates, and go or no-go posture.

Layer 2: Sequencing and Scope Clarity

• order-of-operations enforces dependency-correct task flow.
• requirement-clarifier turns ambiguous requests into explicit acceptance contracts.

Layer 3: Execution Path

• scripted-command-execution handles deterministic local command workflows.
• pseudo-agentic-automation handles dynamic browser or GUI runtime loops.

Layer 4: Risk and Test Integrity

• regression-prevention classifies risk and controls release confidence.
• effective-testing-methods maps code changes to unit and Playwright coverage updates.

Layer 5: Governance and Intent Assurance

• governance-enforcement executes validators and CI policy checks.
• semantic-policy-audit verifies intent-level compliance beyond string or schema checks.
• thoroughly-rate-review applies weighted quality scoring when evaluation is requested.

Layer 6: Documentation and Continuity

• doc-maintenance, file-maintenance, file-structure-optimization, user-instructions-tracker, and history-indexing preserve long-horizon accuracy and traceability.

Layer 7: Product and UX Value

• thoughtful-approach, ui-design-skills, and ui-spatial-canvas drive user-value decisions and coherent interaction architecture.

The canonical cross-skill routing source is /skills/docs/skill-index.md, with execution scenarios in SKILL-MAP.md.

Governance System

Governance is not optional in risky or multi-step work.

Key controls:

1. startup declaration requires explicit skills_in_use, rationale, and execution order
2. validator checks enforce policy snippets, startup declaration integrity, skill catalog sync, and artifact pairing
3. ordering sync check fails if SKILL-MAP.md and docs/skill-index.md drift
4. CI entrypoint enforces governed scope and requires governance artifacts when governed files change
5. regression tests cover governance validator behavior

Primary governance tooling:

• skill-governance/scripts/generate_governance_artifact.py
• skill-governance/scripts/validate_governance_artifact.py
• skill-governance/scripts/validate_skill_policy.py
• skill-governance/scripts/validate_skill_order_sync.py
• skill-governance/scripts/enforce_governance_ci.py
• docs/ci/skills-governance-ci.yml (workflow template)

Prioritization and Decision Order

Default baseline:

1. token-reduction
2. order-of-operations
3. doc-maintenance when behavior, workflow, or policy changes

Conditional priorities:

1. choose governance first for risky, ambiguous, or multi-skill tasks
2. choose command execution mode (scripted-command-execution or pseudo-agentic-automation)
3. apply risk and testing controls before release recommendation
4. enforce documentation and instruction tracker updates before completion

Conflict resolution order:

1. policy (skill-governance)
2. sequence (order-of-operations)
3. risk (regression-prevention)
4. test design (effective-testing-methods)
5. execution (scripted-command-execution)

Intended Outcome

This skillpack is intended to produce:

1. consistent decisions on process strictness and gates
2. fewer regressions on non-trivial changes
3. auditable validation evidence for release readiness
4. tighter alignment between user intent and execution behavior
5. durable documentation and instruction-state continuity
6. faster task completion without reducing safety standards

Maintenance Notes

When adding or changing skills:

1. update SKILL-MAP.md and docs/skill-index.md in the same change
2. update related skill cross-links
3. run governance validators and tests
4. update user-instructions.md with directive and fulfillment evidence