build
Skill0 starsChanged 3 months agoMITView on GitHub
Career Alpha Proof-of-Work 项目构建器。用于把趋势或职业楔子变成 2–7 天可以真正完成、展示和验证的小型项目;当用户问应该做什么项目、如何快速补齐某能力证据、如何把新趋势变成作品集,或需要一个可交给 Codex / Claude Code 执行的项目任务书时使用。
156 other projects use a version of this.
Adopted by
Most recently taken up by
- jskswamy/skillstheir own version, first seen Nov 8, 2026
- domengabrovsek/agent-configtheir own version, first seen Nov 8, 2026
- domengabrovsek/claudetheir own version, first seen Nov 8, 2026
- lilly-builds/senior-engineer-modetheir own version, first seen Nov 8, 2026
- endqwerty/endstacktheir own version, first seen Nov 8, 2026
- DnsQueen/orcinus-skillstheir own version, first seen Nov 8, 2026
Used with
Projects with it also use backprop, finances and 10 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
JuliusBrusseeSkillbackprop
Bug → spec protocol. When a bug is found or a test fails, trace the cause, decide whether a new §V invariant would catch recurrence, append to §B. This is the one non-obvious thing SDD does that plan-then-execute doesn't. Triggers on test failure, bug report, post-mortem, or explicit user ask.
Also in 4 projects that use this
GerardoRdz96Skillfinances
Use when Gera talks about his money — logging an expense ("I spent X on Y"), checking where he stands this pay-period, closing a period, planning debt payoff, onboarding his financial picture, or asking for budgeting/finance opinions. Servy's financial brain. Triggers — "I spent", "log this expense", "where am I this period", "close the period", "how's my budget", "debt plan", "should I pay off", "finances onboard", "/finances".
Also in 3 projects that use this
cpiprintSkillfavicon
Generate favicons from a source image
Also in 3 projects that use this
madeehahassanSkillGo-Live
Pre-launch checklist. Use before making any app public, sharing a link externally, or announcing a product launch. Adapts to Claude Code, claude.ai, and Cowork.
Also in 3 projects that use this
JuliusBrusseeSkilldeepen
Optional design-improvement pass for when you have spare usage to drain. Finds the shallowest modules in the code the spec touches, researches a deeper design, and proposes refactors that shrink interfaces and hide decisions — behavior held constant, tests green before and after. Proposes §I/§V/§T edits, never silent rewrites. Triggers when the user says "deepen this", "improve the design", "this module feels shallow", "pull complexity down", "use spare budget on the codebase", or invokes /ck:deepen. Leans on the codebase-design skill's deep-module vocabulary when present.
Also in 3 projects that use this
synaptiaiSkillgate
Create a BDSK review_gate artifact. Defines a checkpoint that controls whether generated output may proceed toward acceptance. Use when defining quality, security, or scope gates for execution plans.
Also in 5 projects that use this
Other versions156 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 156
- FalkorDB/FalkorDB★ 10ksame name
- microsoft/microsoft-ui-xaml★ 8.6ksame name
- NexaAI/nexa-sdk★ 8.5ksame name
- hathach/tinyusb★ 7.3ksame name
- parcadei/Continuous-Claude★ 3.9ksame name
- parcadei/Continuous-Claude-v2★ 3.9ksame name
- parcadei/Continuous-Claude-v3★ 3.9ksame name
- pytorch/TensorRT★ 3ksame name
- oxsecurity/megalinter★ 2.6ksame name
- iPlug2/iPlug2★ 2.4ksame name
146 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
The file
Files
/build — Build the Evidence
/build 的目标不是“再做一个项目”,而是把 /wedge 的职业假设转成 一段新的、可验证的事实。
核心原则:
Project exists to prove a claim. The claim does not exist to decorate the project.
1. Read the upstream hypothesis
优先读取:
/wedge选出的方向;- 用户当前证据缺口;
- cheapest test;
- kill criteria;
- 目标岗位 / 公司类型;
- 用户可用时间与工具。
如果用户没有经过 /wedge,也可以直接从目标岗位出发,但必须先问自己:
这个项目完成后,简历里会多出哪一条以前不能诚实说的话?如果答案只是“我也做过一个 X”,项目价值通常偏低。
2. Define ONE primary career claim
项目必须有一个主要主张,例如:
- 能设计 Agent Eval;
- 能系统分析 tool-use failure;
- 能做真实数据 pipeline;
- 能优化 latency / cost 并说明 tradeoff;
- 能把模型能力变成完整 workflow;
- 能读懂并扩展真实开源代码库。
可以产生多个 secondary claims,但 scope 由 primary claim 决定。
Bad
做一个完整 AI SaaS,展示产品、前端、后端、算法、增长、商业化能力。
Better
用一个可复现 benchmark 证明我能设计 Agent Reliability Eval。
3. Choose the right project archetype
读取 Project Archetypes。优先从以下类型选择:
- Benchmark;
- Reliability / failure analysis;
- Vertical workflow;
- Integration;
- Cost / latency optimization;
- Open-source extension;
- Reproduction / replication。
不要默认做“全栈网站”。
如果最有价值的证据来自现有开源生态,应考虑把任务转给 /contributor,而不是重新造一个平行项目。
4. Baseline before intervention
除非项目本质无法比较,否则先定义 baseline。
Baseline 至少说明:
- 当前实现 / naïve approach;
- 任务集 / 数据;
- primary metric;
- dependency / model version;
- 已知限制。
没有 baseline 的“优化”“提升”“更可靠”通常无法安全写进简历。
5. Evaluation before polishing
项目早期就定义:
- success metric;
- secondary metrics;
- sample / task set;
- failure taxonomy;
- 是否需要 repeated trials;
- 哪些变量需要固定;
- 哪些结果会被视为 failure。
先测对,再做漂亮。
如果结果不可量化,也至少要有可验证 outcome,例如:
- merged PR;
- deployed workflow;
- external review;
- user feedback;
- reproduced bug;
- working integration with tests。
6. Scope control
使用 Mission Brief Template。
每个 2–7 天任务必须写:
In scope
只保留直接服务 primary career claim 的内容。
Non-goals
强制列出不会做的东西,例如:
- 不做完整登录系统;
- 不做漂亮 dashboard;
- 不支持所有模型;
- 不做 production-scale infra;
- 不做商业化。
Non-goals 是 Career Alpha 的重要部分,因为时间应该花在证据密度最高的地方。
7. 72h-first execution
默认先给 72h 版本,除非任务天然需要外部等待。
Day 1 — Honest baseline
目标:先产生第一份 raw evidence。
常见动作:
- 建 repo;
- 固定数据 / tasks;
- 跑 baseline;
- 保存原始结果;
- 建 failure taxonomy;
- 写第一版 README。
Exit criterion: baseline 可重复,不靠口头解释。
Day 2 — Intervention
- 实现最小核心改动;
- 跑相同 evaluation;
- 保存 raw outputs;
- 记录 regression;
- 不提前筛掉难看的结果。
Exit criterion: 能比较 before / after,哪怕结果没有提升。
Day 3 — Interviewability
- ablation / failure analysis;
- architecture / decision log;
- README reproduction;
- cleanup;
- demo / screenshots;
- 明确 limitations。
Exit criterion: 另一个工程师能理解“问题—选择—证据—限制”。
8. Definition of Done
Definition of Done 必须可观察。
Bad
- 项目基本完成;
- UI 比较完整;
- 效果不错。
Good
- 50-task eval 从一条命令跑完;
- raw output 保存在
results/; - README 写清 baseline 与 intervention;
- 至少 5 个 failure cases 分类;
- architecture diagram 存在;
- 所有结果能对应到 commit / version;
- 用户反馈有原始记录。
9. Evidence capture plan
在开始前规定需要保存:
- repo URL;
- commit hashes;
- changed files;
- benchmark / CSV;
- logs;
- model / dependency versions;
- screenshots / demo;
- test commands;
- user feedback;
- failed attempts;
- scope / ownership notes。
不要项目做完后再凭记忆补证据。
10. Failure is still evidence
Career Alpha 项目不是比赛作品,结果没有提升也可能有价值。
例如:
Context compression reduced cost 22% but caused a measurable regression on long-horizon tasks.这比无证据地写“optimized context efficiency”更能展示工程判断。
所以默认记录:
- what worked;
- what failed;
- regressions;
- unexpected behavior;
- next experiment。
11. Anti-demo rules
以下默认降级:
- 只换 prompt 的 ChatGPT clone;
- 无数据、无 evaluation 的普通 RAG;
- 照教程逐行复刻;
- 只有漂亮 landing page;
- “多 Agent 系统”但没有任务、metric 或 failure analysis;
- 一个大型平台,却没有任何用户或可复现结果;
- 自动生成大量代码,但用户无法解释关键 decision。
Rescue path
如果用户已经有这种项目,不一定丢掉。尝试增加:
- benchmark;
- real user workflow;
- failure taxonomy;
- deployment;
- external contribution;
- cost / latency / quality tradeoff;
- architecture rationale。
12. Execution prompt for coding agents
当用户要交给 Codex / Claude Code 执行时,输出一个独立执行块,至少包含:
Goal
Career claim to prove
Repository structure
Baseline
Implementation tasks
Evaluation
Definition of Done
Non-goals
Evidence to capture
Validation commands
Do not fabricate resultsCoding agent 可以实现和测试,但不能预先写“提升 30%”之类尚未发生的结果。
13. Default output
Mission Brief
- Career claim to prove
- User / problem
- Why now
- Project archetype
- Baseline
- Intervention
- Evaluation
- Scope
- Non-goals
72h / 7-day Plan
按 milestone + exit criterion 写,不只是 todo list。
Repo / artifact plan
说明哪些文件最终构成证据。
Definition of Done
必须全部可验证。
Evidence Capture
提前列出要保存什么。
Failure modes
预判 3–5 个可能让项目证据价值下降的风险。
Coding-agent prompt
用户需要时输出。
14. Handoff to /proof
按 Cross-skill Handoff Contract 传递:
career_claim:
verified_artifacts:
baseline:
results:
ownership:
failures:
limitations:
planned_but_not_completed:
recommended_next_skill: /proof未完成内容必须保留 PLANNED / IN PROGRESS,不能因为 Mission Brief 写得完整就当作已经做完。
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

