# Codex Goal

> Executes complex, long-running autonomous development tasks with thorough self-auditing and verification loops. Operates relentlessly until the stated engineering goal is fully achieved, documenting all phases and validation results in codex-drive/walkthroughs/.

- Skill: `wwwroot/codex-goal` (Agent Skill)
- Install (CLI): `npx skillmds@latest add wwwroot/codex-goal`
- Raw SKILL.md: https://api.skillmd.com/api/skills/wwwroot/codex-goal/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: wwwroot (https://skillmd.com/u/wwwroot)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/wwwroot/codex-goal

---


# Codex Goal — Autonomous Goal Execution & Auditing Skill

> The autonomous execution engine of AI Codex. Designed for complex, multi-step engineering tasks requiring sustained focus, self-auditing, and verifiable completion.

---

## Overview

`codex-goal` is triggered when a user defines an end-to-end goal (e.g., implementing an entire feature suite, refactoring a legacy subsystem, or completing an engineering roadmap). It does not stop at surface-level code edits; it continuously tests, audits, fixes compiler/test errors, and validates invariants until the goal is fully fulfilled. It outputs a complete execution walkthrough in `codex-drive/walkthroughs/`.

---

## When to Trigger

- User runs `/codex-goal` (e.g., `/codex-goal implement end-to-end OAuth2 login flow with tests`, `/codex-goal complete database migration from SQLite to Postgres`, `/codex-goal eliminate all memory leaks in image pipeline`)
- Tasks intended to run autonomously without requiring constant step-by-step user input
- Multi-phase refactoring or complex greenfield implementations

---

## Execution Workflow

```
 ┌────────────────────────────────────────────────────────┐
 │ 1. DEFINE OBJECTIVE & FORMULATE SUCCESS CRITERIA       │
 │    Break the goal into deterministic, testable checks. │
 └──────────────────────────┬─────────────────────────────┘
                            │
                            ▼
 ┌────────────────────────────────────────────────────────┐
 │ 2. ITERATIVE EXECUTION & IMPLEMENTATION LOOP           │
 │    Write code, update manifests, wire components.      │
 └──────────────────────────┬─────────────────────────────┘
                            │
                            ▼
 ┌────────────────────────────────────────────────────────┐
 │ 3. AUTOMATED VERIFICATION & TEST RUNS                  │
 │    Run linters, compilers, unit tests & type checks.   │
 └──────────────────────────┬─────────────────────────────┘
                            │
                            ▼
 ┌────────────────────────────────────────────────────────┐
 │ 4. SELF-AUDIT & RECOVERY (IF FAILING -> RE-ITERATE)    │
 │    Audit against success criteria. Fix all errors.     │
 └──────────────────────────┬─────────────────────────────┘
                            │ (All Checks Pass)
                            ▼
 ┌────────────────────────────────────────────────────────┐
 │ 5. WRITE COMPLETION WALKTHROUGH IN CODEX-DRIVE         │
 │    Write codex-drive/walkthroughs/YYYY-MM-DD-*.goal.md │
 └────────────────────────────────────────────────────────┘
```

---

## Goal Walkthrough Specification (`codex-drive/walkthroughs/`)

All execution walkthroughs generated by `codex-goal` MUST be Markdown (`.md`) files with exact date-time metadata.

### Filename Format:
`codex-drive/walkthroughs/YYYY-MM-DD-<slug>.goal.md`

### Standard Goal Walkthrough Template:

```markdown
# [Goal Title] Autonomous Execution & Verification Walkthrough

> **Created At**: YYYY-MM-DD HH:MM:SS (Local Time)
> **Active Codex Edition**: [`skills/codex/<edition>/`](file:///...)
> **Status**: GOAL_COMPLETED | CANCELLED
> **Total Iterations**: [e.g., 4 iterations]
> **Total Automated Checks**: [e.g., 18 tests passed, 0 errors]

---

## 1. Goal Definition & Success Criteria
- **Stated Goal**: [Full description of the objective]
- **Success Checklist**:
  - [x] All required domain models and schemas implemented
  - [x] Endpoints / UI components fully integrated
  - [x] Zero compiler errors, zero type errors, zero lint warnings
  - [x] 100% automated test suite passing

---

## 2. Changes Summary by Component

### Component A: `src/domain/`
- `[NEW]` [`src/domain/model.ts`](file:///...) — Core domain invariants
- `[MODIFY]` [`src/domain/service.ts`](file:///...) — Integrated business logic

### Component B: `src/api/`
- `[NEW]` [`src/api/routes.ts`](file:///...) — REST/gRPC endpoints

---

## 3. Verification & Self-Audit Trail

| Iteration | Action Taken | Test Result | Self-Audit Verdict |
|---|---|---|---|
| **1** | Scaffolding models and routes | Compilation succeeded, test failed | [WARNING] Incomplete edge case handling |
| **2** | Added boundary validation | Tests passed, linter warned on unused type | [WARNING] Minor lint issue |
| **3** | Cleaned up types & ran full test suite | All 18 tests passed, zero warnings | [OK] **GOAL FULLY ACHIEVED** |

---

## 4. Final Verification Command Output
```bash
$ npm test
[OK] All 18 test suites passed (1.2s)
```

## 5. Next Actions & Operational Readiness
- Code is verified, tested, and ready for deployment or merge.
```

---

## Response Protocol

When `codex-goal` completes:
1. Provide a direct link to `[View Goal Walkthrough](file:///.../codex-drive/walkthroughs/YYYY-MM-DD-<slug>.goal.md)`.
2. Summarize the major changes and test verification results.
3. Include the completion token `<!-- GOAL_COMPLETE -->` in the response.