Hubbestaiskill.com

Home / All Skills / Workflow & Core Tools / dbs-standard-answer

Workflow & Core Tools · dontbesilent2025/dbskill

dbs-standard-answer

For complex questions, this skill first excavates relevant industries, disciplines, authors, and credible theories, then compares structurally isomorphic historical successes, failures, and counterexamples to extract conditioned standard answers. Use when the user needs both theoretical explanations and historical comparisons, or explicitly asks how an issue was historically solved, whether classic solutions exist, or what the standard answer is. Do not use when the goal is merely finding theoretical grounding for an existing proposition, case study, presentation (PPT), or methodology.

L1 Basic · Source registration onlyV1 Commit-PinnedDual-pass Audit

Author dontbesilent2025GitHub dontbesilent2025/dbskill ↗Stars 10121License CC BY-NC 4.0（非商业授权）：仓库根 LICENSE 为 Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)，Copyright (c) 2026 dontbesilent；GitHub API license.spdx_id = NOASSERTION （CC BY-NC 4.0 非 SPDX 标准文本，GitHub 无法自动识别），不得用于商业用途Commit 8b8e33f1ec

Control levels describe skill capabilities and suggested control measures; they do not represent quality or safety certifications, nor do they dictate usability. Usage decisions rest entirely with users and administrators. Different agent hosts apply varying constraints; do not assume hosts provide automatic protection. Third-party skill copyrights belong to their original authors. Use at your own risk.

1How It Works & Mechanism

Rigidly orchestrate a single response into two consecutive sections (Knowledge & Theory Mining → Historical Isomorphism & Standard Answer), while allowing compression based on problem scale.

skills/dbs-standard-answer/SKILL.md

两个部分是统一的分析视角，不是固定篇幅。简单问题可以合并成简短回答，不展开理论地图、历史案例和类比矩阵；需要完整研究时，第一部分找到的机制、边界和反方解释要成为第二部分选择历史案例的依据。

Simple questions can be consolidated into concise answers; when full research is required, both sections must appear and yield substantive conclusions (stopping after only completing part one is prohibited).

Audit the user's original proposition before conducting research: review five risk categories (absolutism, causal leaps, definitional equivocation, empirical generalization, reverse inference) item-by-item, and issue a four-state verdict.

skills/dbs-standard-answer/SKILL.md

| 因果跳跃 | 材料只展示相关、结果或感受，却被写成原因了吗？ |

Four states: Retain / Narrow / Reconstruct / Abandon; after the verdict, state the "revised criteria" first before entering theoretical search.

Screen theories using a scorecard with threshold scores; eliminate any falling below thresholds and prohibit name-dropping.

skills/dbs-standard-answer/SKILL.md

总分低于 7 分不得成为主理论；`机制匹配 < 2` 或 `来源可靠 < 1` 时直接淘汰。候选评分用于内部判断，用户要求研究过程时再展示。

Dimensions are mechanism fit 0–3, source credibility 0–2, figure recognition 0–2, expressiveness 0–2, boundary clarity 0–1; by default retain 1 primary theory + 0–2 auxiliary theories + at least 1 corrective counter-direction.

In part two, first abstract the problem into a "structural fingerprint" (subject / stage / value structure / scarce resources / core tension / lock-in mechanism / pressure mechanism), then locate isomorphic cases accordingly, avoiding superficial analogies based on celebrities or industries.

skills/dbs-standard-answer/SKILL.md

| 锁定机制 | 哪个必要任务持续占用资源，使解决它的能力无法建立？ |

Finally, formulate a structural proposition ("A {subject} at {stage}, relying on {value or delivery structure}, struggles to invest in {capability building} due to {lock-in mechanism}...") as the retrieval query.

Historical cases must mandatorily cover four evidence roles (recent success case / cross-domain repeated sample / failure case / counterexample or edge case), using failures and counterexamples to test whether mechanisms are overvalued.

skills/dbs-standard-answer/SKILL.md

- `反例或边界案例`：未采用主流机制仍然成功，或揭示解法只在特定条件下有效。

Standardize all cases into 12 fields (original context, real constraints, available alternatives, actual decision, execution cost, subsequent outcome, evidence, evidence status, similarities/differences, analogy validity, extractable mechanism, non-transferable elements).

A "standard answer" is defined as a mechanism with conditions, causal explanations, and invalidation boundaries, written in fixed sentence structures; a single celebrity or adage does not constitute an answer.

skills/dbs-standard-answer/SKILL.md

> 当 `{条件}` 成立时，历史上反复有效的做法是 `{机制}`，因为 `{因果解释}`。当 `{边界}` 出现时，这个答案容易失效。

Answers fall into three tiers: mature consensus / conditional answer / no consensus yet; tool names, celebrity habits, or aphorisms alone cannot serve as standard answers.

Tier sources + stop rules to control "textual research bloat": stop once evidence is sufficient; mark as unverified and halt if the same fact fails verification twice consecutively.

skills/dbs-standard-answer/SKILL.md

同一事实连续 2 次无法核实，标注「尚未核实」并停止。

Five priority levels for sources: original works/papers/archives → universities/academic societies/government authoritative publishers → peer-reviewed papers → high-quality secondary interpretations → encyclopedias and quote sites (for lead discovery only).

Web browsing is conditionally triggered and subject to user veto: answer using available materials by default, searching only when external facts substantially alter conclusions or when explicitly requested by the user.

skills/dbs-standard-answer/SKILL.md

只有外部事实会实质改变结论，或用户明确要求核实、来源、最新信息时，才联网；用户要求不联网时，禁止联网。

What is happening here: Web browsing is demoted from a default action to a controlled action to avoid searching merely for academic pretension; actual search is executed by the host tool, and the skill directory contains no endpoints or clients.

The depth of realistic advice is determined by user request, with output controlled by a 5-tier "stopping point" table to prevent over-delivery.

skills/dbs-standard-answer/SKILL.md

| 「先定位」「先研究计划」「暂不解决」 | 第一部分给出知识定位；第二部分给出结构定位、搜索假设和证据计划 |

Corresponding tiers: general inquiry → 1–3 verifiable actions; past research → case + matrix + standard answer; integrate present → append real-world actions; full plan → handover-ready constraints, mechanisms, and boundaries.

2Core Capabilities

01Deliver direct answers first (prohibiting "please provide theories first" as a deliverable)

02Proposition auditing (absolutism / causal leaps / definitional equivocation / empirical generalization / reverse inference) and four-state verdicts

03Cross-domain knowledge mapping: classifying fields / industries / authors / theories into Support / Supplement / Corrective-Refutation

04Theory and figure verification: figure positioning phrasing, Chinese translation rules, theory cards (three tiers: original meaning / source / verification status)

05Theory-to-public-problem mapping table (6–10 directions, filtering down to 3–5 priority directions by frequency / pain point / explanatory power, etc.)

06Short video application: default 3 archetypes (hook / cost / theoretical explanation / minimal action / demonstration / validation / payoff)

07Historical isomorphism research: structural fingerprint → 3–6 candidate problem families → cases across four evidence roles → analogy matrix

08Standard answer tier assessment (mature consensus / conditional answer / no consensus yet) and invalidation boundaries

09Source tiering and stop rules (including "halt after 2 consecutive verification failures")

3External Dependencies

Kind	Dependency / Endpoint
network	宿主联网检索能力（宿主工具，非本 skill 目录内资产；条件触发）

4Recommended Governance & Evidence

Suggested Controls for Safe Operation

For Agent Runtime

Can be auto-loaded

For Administrators

Register source and version

Evidence & Triggers (File Paths & Excerpts)

SKILL.md: Pure prompt instructions and rule text; no script execution or external impact.

Audit Notes: According to the Batch 4 final criteria where the "evaluation subject = skill directory itself": The directory contains only SKILL.md + agents/openai.yaml, with no scripts, no endpoints, no credential access, and no local disk writes (only producing in-chat Markdown). Although the workflow allows falling back to host web browsing under specific conditions, that relies on the host tool, and the skill specifies no concrete backend or domain (endpoint count is 0). Furthermore, the source code hardcodes the default state and user veto power ("prohibit web browsing when user requests not to browse"). Under the rule that "endpoints appearing only in prohibited/optional clauses → green," the tier is not escalated and is noted only in reason/risks.

5Second-Pass Verification

[ok] 双部分编排与输出顺序 — 第一部分/第二部分标题与"最终输出顺序"5 条（直接答案 → 第一部分 → 第二部分 → 综合结论 → 资料来源）均在源码中。
[ok] 理论评分门槛数值准确 — "总分低于 7 分不得成为主理论"与"`机制匹配 < 2` 或 `来源可靠 < 1` 时直接淘汰"逐字命中。
[ok] 四类历史证据角色齐全 — 最近成功案例/跨域重复样本/失败案例/反例或边界案例四条均在「Step 4：研究历史案例」中。
[ok] 标准答案句式与三档等级 — 条件句式原文存在；"成熟共识／条件性答案／尚无统一答案"三档逐字存在。
[ok] 联网门控与用户否决权 — 原文明确"用户要求不联网时，禁止联网"，并给出 5 级来源优先级与"连续 2 次无法核实…停止"规则。
[ok] 外部依赖清单是否完备 — 全目录扫描无脚本/CLI/API/域名，无第三方包；目录内无任何运行时端點，联网一项已标注为宿主能力、非目录内资产。
[ok] 是否声明了不适用场景（与 theory-grounding 不重叠） — description 末句"只需给现有命题、案例、PPT 或方法论找理论依据时不使用"，与 dbs-theory-grounding 的适用描述形成互补。
[unlocatable] 联网的具体实现与端点 — skill 只以自然语言授权"联网"，未指定任何域名、CLI 或 API；实际检索由宿主工具实现，无法在源码内定位具体端点与请求形态。

6Verdict & Scenarios

Redefines "standard answer" as a mechanism equipped with conditions and failure boundaries, rather than a celebrity quote or isolated story.

Anti-confirmation bias design: enforces simultaneous study of failure cases and counterexamples to test whether mechanisms are overstated.

Auditable research process: 5 source tiers, clear distinction between verified facts / researcher interpretations / current inferences / unverified items, with clickable sources required.

Resistant to "textual research bloat": minimal sufficient search + stop when evidence is sufficient + mark as unverified after 2 consecutive verification failures, accompanied by depth control tables to avoid over-delivery.

Multi-format delivery: the same research can be adapted into short video / PPT / course formats (defaults to 3 short video archetypes) for practical content creation.

Best for: Suitable for complex questions requiring both "theoretical explanation" and "historical comparison": finding theoretical and historical grounding for self-developed methodologies, determining whether a tactic has been validated in other industries, and preparing citation-backed argumentative frameworks for courses, PPTs, or long-form essays. Also suitable for scenarios explicitly requesting to "investigate how this was historically resolved and whether classic solutions exist."

Not for: Not suitable for tasks that merely require finding theoretical support for existing propositions, cases, PPTs, or methodologies (which are explicitly covered by dbs-theory-grounding); not suitable for scenarios requiring immediate answers for quick decision-making without time to read a two-part study; not suitable for quantitative conclusions requiring empirical data or lab experiments; when the user requests not to browse the web, facts requiring external verification will remain marked as "unverified."

Install Agent-Ready

① Verified Mirror 2026-09-15

Option A · Direct DownloadDownload dbs-standard-answer.tar.gz

sha256: 4152c87117c20590…

Option B · Copy Guide for Agent

Agents automatically download and verify mirrors from this JSON guide.

② Upstream GitHub

GitHub Repo ↗

Pinned to commit 8b8e33f1ec.