goal-verification

A guide for designing goals and completion checks in an automated loop, where an agent repeatedly performs work. It turns a vague objective into a measurable target and a clear test for knowing when the work is done.

In plain words
What is it for?
Use it when designing or debugging an agent loop, especially when it will not stop, produces uneven quality, or has no clear definition of finished work.
Why use it?
It helps prevent inconsistent results and loops that keep running because nobody defined completion. It replaces subjective checks such as “make it good” with verifiable conditions.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/kangarooking/loop-engineering-skill/goal-verification
Any agent
npx skills add kangarooking/loop-engineering-skill --skill goal-verification
Clone the repo
git clone --depth 1 https://github.com/kangarooking/loop-engineering-skill

Made for: Claude Code, Codex.

Per session 105 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,600 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00105 $0.01600
Opus 5 $0.00053 $0.00800
Sonnet 5 $0.00021 $0.00320
Haiku 4.5 $0.00011 $0.00160

Measured 2d ago against content hash abc1843420d1, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

goal-verification scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

goal-verification/SKILL.md · 125 lines

How it starts

The opening of the file, as written. The whole thing — 125 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Source Metadata

Original cangjie-skill frontmatter from the distillation run:

name: goal-verification
description: |
  设计 Loop 的 Goal 和 Verification 环节: 将模糊目标转化为可验证的停止条件。
  当用户发现 loop 产出质量不稳定、循环无法停止、或"不知道什么时候算做完"时使用。
  不适用于: 目标已经非常清晰可量的任务、或非循环场景。
  关键 trigger: "loop 停不下来"、"产出质量不稳定"、"怎么定义完成"、"goal 怎么写"。
source_book: "Loop Engineering 视频合集"
source_chapter: 视频1 (Adam Gillock) / 视频4 (Idoos Money)
tags: [goal-design, verification, stop-condition, quality]
related_skills: [loop-three-elements, maker-checker, loop-build-path]

Goal 可验证化 — 循环设计的质量杠杆点

R — Reading (原文)

"There's really two most important pillars: the goal (objective, not subjective) and then verification. How does the agent know what that stop condition is?" — Adam Gillock (视频1)

"A loop is only going to be as good as its done check, as the done criteria." — Adam Gillock (视频1)

I — Interpretation (自述)

循环的质量上限 = 其验证环节的质量上限。设计 Loop 时,必须回答两个问题:

  1. Goal (目标): 循环要达成什么? 必须是客观可验证的 — 不是"做好",而是"达到 X 指标"。
  2. Verification (验证): 如何判断目标已达成? 必须有可执行的检查步骤

好 Goal vs 坏 Goal:

  • ❌ "直到你满意" — 主观,不可验证
  • ❌ "做好这个功能" — 模糊,无法判断
  • ✅ "凑够 5 条数据" — 客观,可数
  • ✅ "平均分 ≥ 9 或最多 8 轮" — 客观 + 硬停

验证方式光谱 (从客观到主观):

  • 纯客观: 跑测试套件、数值指标
  • 半客观: 另一个 LLM 评判
  • 主观: 人类在环判断
  • 模糊: 无法明确判断好坏

A1 — Past Application (书中案例)

案例1: 坏 Goal — 缩略图生成 (视频1)

  • Goal: "迭代直到满意" → 主观,导致 27 分钟不可控
  • 教训: 应改为"评分 ≥ 8/10 或最多 5 轮"

案例2: 好 Goal — Abbey Road 复刻 (视频1)

  • Goal: "平均分 ≥ 9 或最多 8 轮" → 客观 + 硬停
  • 虽然结果不完美,但循环在预期内停止

案例3: 验证光谱应用 (视频4)

  • 测试套件 → 最客观,优先用
  • LLM 评判 → 次选,需独立 checker
  • 人类判断 → 最后手段,成本高

A2 — Future Trigger (未来触发)

  1. Loop 无法停止时: "跑了 3 天还没完" → 检查 stop condition 是否过主观
  2. 产出质量不稳定时: "有时好有时坏" → 验证环节可能不可靠
  3. 设计新 Loop 时: 先写 goal 和 verification,再写 action
  4. 调试 Loop 时: 诊断"为什么产出差" → 先看 goal 是否可验证

语言信号: "loop 停不下来"、"怎么定义完成"、"goal 怎么写"、"产出质量不稳定"

与相邻 skill 的区别:

  • loop-three-elements: 三要素的整体框架 (本 skill 专注 stop condition 的设计)
  • maker-checker: 验证环节的具体实现 (本 skill 是验证的设计原则)
  • loop-worthiness-test: 判断要不要做 loop (本 skill 是决定后如何设计)

Read the full file on GitHub · 125 lines

Files

What ships with it

2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 125 lines · 105 tokens per session scan A abc1843420d1

Subscribe to this mod's changes

goal-verification is a skill published in the GitHub repository kangarooking/loop-engineering-skill (23 stars, last pushed 2mo ago), licensed MIT. It adds 105 tokens to every session and 1,600 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.

obra/superpowers · 21 tokens

brainstorming

You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.

obra/superpowers · 37 tokens

auto-perf-optimize

Run agent-driven VS Code performance or memory investigations. Use when asked to launch Code OSS, automate a VS Code scenario, run the Chat memory smoke runner, capture renderer heap snapshots, take workflow screenshots, compare run summaries, or drive a repeatable scenario before heap-snapshot analysis.

microsoft/vscode · 62 tokens

chat-perf

Run chat perf benchmarks and memory leak checks against the local dev build or any published VS Code version. Use when investigating chat rendering regressions, validating perf-sensitive changes to chat UI, or checking for memory leaks in the chat response pipeline.

microsoft/vscode · 51 tokens

chat-pet-sprite-creation

Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.

microsoft/vscode · 53 tokens

cpu-profile-analysis

Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…

microsoft/vscode · 71 tokens