AgentStack
SKILL verified MIT Self-run

Ccdawn Goal Loop

skill-ccdawn-codex-skills-ccdawn-goal-loop · by CCDawn

Use when the user explicitly wants a persistent iterative goal contract with bounded inputs/tools, strict constraints, success evidence, repeated next-step selection, and stop conditions, and no more specific CCDawn owner already controls the work; do not use for ordinary one-turn or finite implementation requests.

No reviews yet
0 installs
2 views
0.0% view→install

Install

$ agentstack add skill-ccdawn-codex-skills-ccdawn-goal-loop

✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

Are you the author of Ccdawn Goal Loop? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Goal Loop

这个 skill 用来把一个目标变成可执行、可验证、可停止的工作回路。

它不是泛泛的 brainstorming,也不是 BRT 的替代品。它关注的是:目标是否清楚、证据是否足够、约束是否被守住、下一步是否值得做、以及什么时候必须停下来报告。

BRT interface

  • Context Boundary: 当前目标、可用输入/工具/文件范围、明确约束、已获得证据和本轮迭代对象。
  • Output Contract: goal contract、当前行动、advance/iterate/stop 决策、证据和 Route Out。
  • Allowed Action: 只执行 goal contract 允许的最小下一步;写入、联网、安装、删除、迁移、发布或权限动作必须在 contract 中显式允许。
  • Success Evidence: 当前行动产出能满足或推进 Evidence 字段的可检查证据。
  • Stop Condition: 约束会被突破、没有有效下一步、缺少必须信息/权限、连续失败无法产生新证据、或用户要求暂停。
  • Route Out: 继续 goal loop、回到 ccdawn-brt 重新对齐目标、进入具体 owner skill、ccdawn-completion-summary,或 BLOCKED。

统一调用契约

  • 只处理 BRT interface 范围;不匹配时回 ccdawn-brt 或更具体 owner,复合任务不吞其他 owner。
  • 用户可见内容默认中文,完成只报状态、产出、证据和剩余风险;代码、命令、路径、错误原文、API/协议、skill 名和枚举保留原样;Route Out 仅以 BRT interface 为准,末行写 下一步建议:

Goal contract

先把下面 6 个字段补齐:

  • Goal:Codex 最终要完成什么
  • Evidence:什么证据能确认结果有效
  • Constraints:哪些限制条件不能被破坏
  • Allowed scope:允许使用的输入、工具、文件范围或操作边界
  • Next-step rule:每一轮迭代后如何判断下一步最优行动
  • Blocker protocol:遇到阻塞或无路可走时如何停止并汇报

映射到 BRT 时:

  • Goal -> Desired Outcome;
  • Evidence -> Success Evidence;
  • Constraints + Allowed scope -> Allowed Actions / Out of Scope;
  • Next-step rule -> Route Out 和下一轮选择规则;
  • Blocker protocol -> Stop Condition 和 BLOCKED 输出。

Operating rules

  1. 如果字段缺失,并且会影响执行路径,先只问最关键的一个。
  2. 在动手前,先复述一遍填好的 goal contract。
  3. 每一轮只做能最大化新证据、同时不突破约束的最小下一步。
  4. 每轮结束时,明确判断是 advanceiterate 还是 stop
  5. 如果已经阻塞,或者没有有效路径可以继续,必须停止,不要硬推进。

Stop rule

一旦停止,必须报告:

  • 已经尝试过的方法
  • 已获得的证据
  • 当前阻塞点
  • 还需要什么信息、权限或资源才能继续

Output format

响应时使用:

Goal:
Evidence:
Constraints:
Allowed scope:
Next-step rule:
Blocker protocol:
Allowed action:
Current action:
Decision:
Route Out:
Success Evidence:
Stop Condition:
If blocked:
- Tried methods:
- Evidence obtained:
- Current blocker:
- Needed info or permission:

When to use

  • 用户给出一个明确目标,但需要先变成可执行合同
  • 用户要求每轮迭代后根据证据和约束决定下一步
  • 用户要求遇到阻塞时停止并报告
  • 用户需要一个 bounded workflow,而不是开放式探索

See [EXAMPLES.md](EXAMPLES.md) for concrete prompt patterns and stop cases.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet — be the first.

Versions

  • v0.1.0 Imported from the upstream source.