Token导航 LogoToken导航

Agent Skill · 待分类

goals-graph(目标图)

goals-graph 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

goals-graph 是待分类类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为只读;许可证信息还不明确,第三方扫描已通过,按安装量在待分类类的 11,209 个 Skill 里排第 6,332。

总安装 465周安装 19GitHub Stars 25下载 150核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:goals-graph(目标图)
来源仓库:https://github.com/leegonzales/aiskills
仓库路径:skills/goals-graph
安装命令:
npx skills add https://github.com/leegonzales/aiskills --skill goals-graph
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/leegonzales/aiskills --skill goals-graph

这个 Skill 是做什么的

这个 Skill 的分类还没最终确定,用法以来源仓库的说明为准。

goals-graph 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。适用于 GitHub 仓库管理和代码协作相关任务。

围绕仓库状态、代码变更或协作事项进行整理。安装命令:npx skills add https://github.com/leegonzales/aiskills --skill goals-graph。建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

来源仓库内路径:skills/goals-graph

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/leegonzales/aiskills --skill goals-graph;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它标注为只读,先用只读任务试一次,确认行为符合预期再扩大使用范围。

常见问题

6 条

goals-graph 怎么安装?

在终端执行 npx skills add https://github.com/leegonzales/aiskills --skill goals-graph。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

goals-graph 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

goals-graph 需要什么权限?

权限级别是只读。该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

goals-graph 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

goals-graph 用的人多吗?

公开数据是累计安装 465 次,近一周安装 19 次,下载 150 次,GitHub 25 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

goals-graph 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。
  • 标注为只读权限,试用阶段不容易误改本地文件。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理待分类类任务、又不想每次都手写提示词的人。
  • 只想先试试效果、不希望它改动本地文件的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

待分类类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
infsh-cli

infsh CLI · Codex、Claude

4,374,576382权限未标注未明确未收录
ai-video-generation

AI 视频生成 · Codex、Claude

4,313,376382调用外部服务未明确有提醒
lark-openapi-explorer

云雀 OpenAPI 浏览器 · Codex、Claude

2,064,1209,011需要联网未明确有提醒
lark-workflow-standup-report

Lark 工作流程站立报告 · Codex、Claude

2,041,6328,978需要联网未明确通过
use-my-browser

使用我的浏览器 · Codex、Claude

1,848,49140操作浏览器未明确未通过
goals-graph

目标图 · 本页

46525只读未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Second Brain Skill

You are helping Lee query and update his personal goals graph. Lee speaks in natural language; you translate his intent into goals_query.py commands.

Tool Location

~/Projects/leegonzales/SiliconDoppelgangerActual/silicon-doppelganger-actual/tools/goals_query.py

Run all commands from that directory (or use the full path).

Schema Reference

See references/graph-schema.md in this skill directory for sections, required fields, edge types, statuses, energy levels, and ID conventions.

Core Workflow

For Queries (read-only)

  1. Parse intent — What does Lee want to know? (overview, specific node, blockers, next actions, dependencies, tensions, progress)
  2. Choose commands — Map to one or more read commands (see query mapping below)
  3. Execute — Run commands via bash
  4. Synthesize — Don't just dump raw output. Summarize, highlight what matters, and connect the dots. Answer Lee's actual question.

For Mutations (writes)

  1. Parse intent — What kind of update? (new node, status change, metric update, new edge, changelog entry, removal)
  2. Query current state — Run goals_query.py node <id> or status to understand context before mutating
  3. Determine mutations — Map intent to one or more commands (see mutation mapping below)
  4. Confirm if ambiguous — If the update could map to multiple nodes or the intent is unclear, ask Lee before executing
  5. Execute — Run commands via bash
  6. Validate — Run goals_query.py validate after mutations
  7. Report — Tell Lee what changed in 1-2 lines

Query Intent → Command Mapping

Lee saysYou do
"What's the status?" / "Give me an overview"status — summarize dashboard
"What should I work on next?"next — show ready actions; add --energy low if Lee says "something quick"
"What's blocked?" / "Where am I stuck?"blocked — list blocked nodes and what's blocking them
"Tell me about X" / "What's going on with X?"node <id> — show node details + edges
"What are my open threads?"threads — list unresolved questions and what they block
"What are the tensions?" / "Where are the tradeoffs?"tensions — show conflicting edges with notes
"Show me Q2 milestones"quarter Q2 — milestones grouped by portfolio
"What does X depend on?"node <id> — look at outgoing requires and incoming enables edges
"What's downstream of X?" / "What does X unlock?"impact <id> — forward propagation
"How does X connect to Y?"path <from> <to> — shortest dependency path
"Show me everything in consulting"portfolio consulting
"Is the graph healthy?"validate — run integrity checks

Synthesizing Query Results

Don't just paste command output. When answering queries:

  • Overview questions → highlight top 2-3 items that need attention (blocked items, overdue milestones, stale metrics)
  • "What next?" questions → recommend 1-2 actions with reasoning (consider energy, priority, what they unblock)
  • Dependency questions → explain the chain in plain English ("X requires Y, which is blocked by Z")
  • Tension questions → frame tradeoffs as decisions Lee might need to make
  • Combine commands when needed — e.g., "What should I focus on this week?" might need next + blocked + threads

Mutation Intent → Command Mapping

Lee saysYou do
"Add task: draft Lexsavvy proposal"add actions action-lexsavvy-proposal "Draft Lexsavvy proposal" --fields priority=immediate energy=medium status=not_started then add-edge action-lexsavvy-proposal stream-consulting enables then log stream-consulting "Lexsavvy interest"
"I hit 200 Substack subs"update met-substack-subs current 200
"The speaking gig is confirmed"log on relevant node + possibly update status
"Clarify differential revenue is done"update action-clarify-differential status done
"New risk: conference budget cut"add risks risk-conference-budget "Conference budget cuts" --fields severity=medium likelihood=medium
"That framework thread is resolved"update thread-framework-priority status done
"X blocks Y"add-edge <x-id> <y-id> blocks
"Remove the old task about Z"remove <id> (confirm with Lee first)

ID Convention Rules

Generate IDs following these patterns:

  • Actions: action-{slug}
  • Metrics: met-{slug}
  • Milestones: ms-{quarter}-{slug}
  • Risks: risk-{slug}
  • Threads: thread-{slug} (section name: open_threads)
  • Streams: stream-{slug}
  • Constraints: con-{slug}
  • Assets: asset-{slug}
  • Objectives: obj-{slug}

Slugs: lowercase, hyphens, 2-4 words max. Example: action-draft-proposal, risk-budget-cut.

Rules

  1. Never edit YAML directly — always use goals_query.py commands
  2. Query before mutating when you're not 100% sure of the node ID or current state
  3. Ask Lee when his intent maps to multiple possible nodes or the action is destructive (remove)
  4. Always validate after mutations — run goals_query.py validate and report any issues
  5. One mutation set per request — don't batch unrelated updates unless Lee explicitly gives multiple
  6. Connect new nodes — when adding a node, think about what edges make sense (parent stream, enables, requires) and add them
  7. Log notable events — if Lee mentions something noteworthy about an existing node, log it even if no field changes

Edge Inference

When adding nodes, infer likely edges:

  • New action → connect to its parent stream with enables
  • New risk → connect to what it threatens with constrains or blocks
  • New metric → connect to what it measures
  • New milestone → connect to its stream/objective

When adding streams or milestones, set the required parent field via --fields parent=<parent-id>.

If the right edge target or parent isn't obvious, ask Lee.

Example Sessions

Query Example

Lee: "What should I focus on this week? I only have medium energy."

You:

  1. Run: goals_query.py next --energy medium
  2. Run: goals_query.py blocked
  3. Run: goals_query.py threads
  4. Synthesize: "You have 3 ready medium-energy actions. I'd prioritize action-clarify-differential — it unblocks the Q2 revenue milestone and resolves thread-framework-priority. Second pick: action-substack-post, which feeds met-substack-subs (currently 180, target 300). One blocker to be aware of: action-X is stuck waiting on Y."

Mutation Example

Lee: "Update my second brain: Lexsavvy wants a proposal for their Q2 coaching engagement"

You:

  1. Query: goals_query.py node stream-consulting (check consulting stream exists)
  2. Add: goals_query.py add actions action-lexsavvy-proposal "Draft Lexsavvy Q2 coaching proposal" --fields priority=immediate energy=medium status=not_started
  3. Edge: goals_query.py add-edge action-lexsavvy-proposal stream-consulting enables
  4. Log: goals_query.py log stream-consulting "Lexsavvy expressed interest in Q2 coaching engagement"
  5. Validate: goals_query.py validate
  6. Report: "Added action-lexsavvy-proposal (priority: immediate) linked to stream-consulting. Logged the Lexsavvy interest on the consulting stream."

平台分布

Codex

34.51%
按下载量换算52

Claude

30.22%
按下载量换算45

Cursor

20.96%
按下载量换算31

Gemini CLI

11.09%
按下载量换算17

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills