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.
npx agentmods add skills/taichuy/1flowbase/problem-framingnpx skills add taichuy/1flowbase --skill problem-framinggit clone --depth 1 https://github.com/taichuy/1flowbaseWhat 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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00171 | $0.01999 |
| Opus 5 | $0.00086 | $0.01000 |
| Sonnet 5 | $0.00034 | $0.00400 |
| Haiku 4.5 | $0.00017 | $0.00200 |
Grade A, and why
problem-framing 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.
How it starts
The opening of the file, as written. The whole thing — 131 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Problem Framing
Outcome
把请求收敛成可决策、可执行、可验收的结果,不替实现者规定完整路径。完成时,现状有证据,结果可观察,范围、owner 与授权闭合,验证足以结算风险,并有唯一建议和停止条件。
本 Skill 只形成决策,不修改产品代码、测试、migration、schema 或运行时行为。
Trigger Boundary
- 把“为什么新增和编辑不一致”“为什么某能力看不见 / 不能改”“接口或状态应由谁拥有”“这是缺陷还是设计边界”等请求视为需求对齐,即使用户使用“看看原因”“诊断一下”等查询措辞。
- 只有答案不会改变产品行为、交互、contract、owner、成功标准或后续改动方向时,才把请求视为可跳过的事实查询。
- 在完成本 Skill 的决策输出前,不进入 implementation / QA Skill,也不让专项代码探索替代需求分析;需要证据时只获取会改变方向的最小证据。
Reasoning Catalysts
先推理后结论:先定义理想结果,再用第一性原理拆出事实、隐藏因果、硬约束与失败模式;用奥卡姆剃刀选择足以解释证据的最小机制;先升温发散真实方向,再降温收敛唯一建议,最终通俗易懂但不牺牲准确性。
这里的“先推理”指先输出可核验的现状与需求分析摘要,再给方向和最终建议;不以“结论 / 建议”开头,也不展示内部思维链。
优先用高信息关系、反例或最小案例催化,不堆模型已知常识和同义说明。催化词只改变搜索方向,不替代证据、领域精度或硬边界,也不作为口号复述。
Architecture Catalysts
Deep Modules / Information Hiding:公共接口只暴露调用方决策所需的最小充分信息;状态判断、协议细节与兼容分支留在内部。Conservation of Complexity + Requisite Variety(Tesler's Law/Ashby's Law):必要复杂度不能消失;由拥有足够状态与动作空间的语义 owner 吸收。Observability × Controllability ⇒ Ownership:看不见相关状态或不能控制其转移的模块,不拥有该复杂度。Proven Mechanisms over Ad-hoc Rules:优先成熟数学关系、算法、数据结构、状态机、约束与调度机制,不用临时规则堆叠代替。
用以下 complexity placement heuristic 选择 owner;这是本 Skill 的架构判定式,不是经典控制论原公式:
owner*(x) =
argmin_m [C_leak(m) + C_coordination(m) + C_failure(m)]
subject to:
SourceOfTruth_m(x)
∧ Observable_m(x)
∧ Controllable_m(x)
∧ Variety_m ≥ Variety_x
C_leak 是泄漏给调用方的兼容、分支与隐式约定;C_coordination 是跨 owner 协调成本;C_failure 是复杂度错置造成的失败成本。
Decision Field
把请求看作受约束决策;用关系筛选内容,不机械复述检查字段,但最终答复必须遵守 Response Contract:
证据 -> 现状 -> 与目标的差距 -> 可观察成功标准
source of truth / owner -> 必要复杂度
授权 / contract -> 可行方向
失败风险 -> 验证强度
潜在决策变化 × 影响 > 获取成本 -> 新证据
维持以下守恒关系:
- 用户描述提供线索,结论强度不超过证据;安全、数据、权限与已确认 contract 是硬边界,工作偏好只改变方向权重。
- 方案范围不超过授权与非目标;新增范围同时产生成功标准、owner、证据责任和资源边界。
- 只处理会改变可行域或推荐的未知;其他缺口使用有界假设。下一步不能减少决策残差时停止。
- 后端是 contract 与状态唯一数据来源;前端不承担输出兼容,接口字段保持后端 DTO / 领域语义原名。
Control Loop
[证据与结果差距] -> [三个真实方向] -> [唯一建议] -> [用户决策]
^ |
+---- 边界、语义或授权发生变化 --------+
What ships with it
7 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.
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.
- 2d ago First seen · 131 lines · 171 tokens per session scan A 18ef67b3fc03
problem-framing is a skill published in the GitHub repository taichuy/1flowbase (259 stars, last pushed 3d ago), licensed Apache-2.0. It adds 171 tokens to every session and 1,999 once invoked, about $0.0009 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.
Other skills, from other repositories
sync-cliproxy-core
Use when asked to 同步、更新、升级或审计 CLIProxyAPI、cliproxy、translator core、provider adapters、Antigravity 请求/响应转换、internal/protocol/cliproxy 转换快照,刷新上游 commit,或审查一次核心与渠道适配器的原子同步结果。.
ccload-release
用于发布 ccLoad 新版本,自动提交未提交改动并推送本地领先的 master,按固定版本通道计算并发布 Tag,等待 GitHub Actions,以及验证 GitHub Release 和对应通道的容器镜像。Beta 固定沿用最近稳定版的主版本和次版本;只有显式 stable 发布才允许修改次版本。.
10router-chat
Chat / code generation via 10Router using OpenAI /v1/chat/completions or Anthropic /v1/messages format with streaming + auto-fallback combos. Use when the user wants to ask an LLM, generate code, summarize text, or run prompts through 10Router.
10router
Entry point for 10Router — local/remote AI gateway with OpenAI-compatible REST for chat, image, TTS, embeddings, web search, web fetch. Use when the user mentions 10Router, TENROUTERURL, or wants AI without writing provider boilerplate. This skill covers setup + indexes capability skills; fetch the relevant capability…
unified-ai-gateway
Operate and evaluate Unified AI System through twelve governed MCP tools, including provider-free prompt enhancement, while preserving fake-provider, authorization, and evidence boundaries.
10router-image
Generate images via 10Router /v1/images/generations using OpenAI / Gemini Imagen / DALL-E / FLUX / MiniMax / SDWebUI / ComfyUI / Codex models. Use when the user wants to create, generate, draw, or render an image, picture, or text-to-image (txt2img).