科研 Skill
论点对抗性压力测试
stress-testing-argument
Adversarially tests a thesis against counterexamples, rival explanations, and scale violations. Use before finalizing an argument to probe its weak points.
SKILL.md
Stress-Testing an Argument
Purpose
Identify the strongest informed objections to an argument, rank their severity, and specify repairs without reducing the thesis to a trivial claim. The skill challenges evidence, warrants, concepts, scope, method, ethics, and novelty while preserving fair representation.
Contract
Accepts
- A thesis, enabling claims, argument map, evidence inventory, and scope boundary.
- Optional rival theories, disciplinary criteria, reviewer comments, or venue expectations.
Requires
- A sufficiently explicit argument to reconstruct charitably.
- Access to the evidence and warrants being tested.
- Separation between verified criticism and hypothetical reviewer concerns.
Produces
- A steelmanned argument summary.
- Ranked
Objectionrecords. - Alternative explanations and counterexamples.
- Recommended responses: rebut, concede, narrow, redesign, or defer.
- A stress-test gate report.
May produce
- A revised thesis or scope boundary for researcher approval.
- A fatal-flaw report.
FAILwhen the argument cannot survive without abandoning its central contribution.
Fails when
- The argument or evidence is unavailable.
- Critique would depend on invented sources, consensus, or reviewer preferences.
- The user requests abuse, humiliation, or predetermined rejection rather than analysis.
- Ethical or legal issues require specialist review beyond the skill’s competence.
Guarantees
- The argument will be represented in its strongest fair form before criticism.
- Objections will identify a target claim, reasoning, severity, and repair path.
- Residual risks will remain visible.
Does not guarantee
- Reviewer acceptance, truth, novelty, or freedom from all objections.
- That the proposed repair preserves every ambition of the original thesis.
When to use
Use after an argument map exists, before submission, during major revision, or when a new concept appears to explain too much. Use when evidence may support rival accounts, when a local interpretation expands into a historical or ontological claim, or when novelty depends on an uncertain genealogy.
Do not use as performative hostility. “Reviewer #2” theater that rewards harshness over accuracy is an anti-pattern.
Inputs required
Collect:
- Thesis and enabling claims.
- Argument map with warrants.
- Evidence and citation status.
- Scope boundary.
- Concept definitions and lineage.
- Known alternatives and negative cases.
- Ethical, political, or disciplinary constraints.
If the source base is missing, critique only the formal structure and mark evidentiary judgment unknown.
Procedure
1. Steelman the argument
Restate the thesis, claim sequence, evidence, and stakes in the strongest form the supplied materials permit. Ask for researcher confirmation when reconstruction changes emphasis.
2. Obtain steelman confirmation
Present the reconstruction to the researcher before adversarial testing. Ask whether it is the strongest version the researcher is prepared to defend and identify any omitted qualification, evidence, or conceptual commitment. Do not proceed as though silence were approval.
When interactive confirmation is unavailable, mark the reconstruction researcher confirmation required and limit subsequent objections to provisional status. An agent-generated steelman is not evidence that the author's strongest argument has been captured.
3. Prioritize objection dimensions
Choose objection dimensions according to the argument type rather than treating every category as equally important.
- interpretive argument: prioritize object fidelity, warrants, rival readings, and scale;
- historical or genealogical argument: prioritize chronology, transmission evidence, archive coverage, and alternative causal accounts;
- conceptual argument: prioritize definition, boundary, category error, internal consistency, and counterexample;
- comparative argument: prioritize comparability, asymmetry, selection criteria, and translation;
- normative argument: prioritize stated values, affected parties, feasibility, and ethical consequences.
Record why the selected dimensions receive priority.
4. Build an objection matrix
Test at least these dimensions:
- evidence adequacy;
- warrant validity;
- concept definition;
- category and level consistency;
- scope and representativeness;
- historical or genealogical support;
- rival explanation;
- counterexample and negative case;
- normative assumption;
- ethical or political consequence;
- false or inflated novelty.
5. Test inferential transformations
Look for prohibited movement:
interpretation → fact
correlation → causation
local reading → universal claim
similarity → influence
source mention → source support
new term → new concept
6. Generate serious alternatives
Construct explanations that account for the same evidence with fewer assumptions or different causal levels. State what additional evidence would discriminate among them.
7. Search for counterexamples
Identify cases inside and outside the corpus that would weaken the claim. Do not invent cases; use supplied or verified examples, or mark candidates as to verify.
8. Audit scope
Compare every major conclusion with the scope boundary. Flag movement across text, corpus, medium, period, geography, institution, or ontology without an explicit bridge.
9. Audit concepts
Check equivocation, circular definition, excessive extension, hidden normativity, and collapse of neighboring terms. A concept that explains every outcome is not discriminating enough.
10. Rank severity
Use:
- Fatal: central thesis cannot stand under current evidence or logic.
- Major: substantial revision, new evidence, or re-scoping required.
- Moderate: argument remains viable but a section or qualification is needed.
- Minor: clarity, terminology, or presentation issue with limited substantive effect.
Also estimate likelihood that an informed reviewer would raise the issue, while marking this as inference rather than fact.
11. Select response strategy
For each objection recommend:
- rebut;
- concede;
- narrow;
- redesign;
- defer as limitation;
- seek specialist review.
Do not recommend rhetorical evasion.
12. Re-test the repaired thesis
When a repair is proposed, test whether it becomes trivial, circular, or inconsistent with the original question. Record what intellectual cost the repair imposes.
13. Issue the gate report
Return PASS only when no fatal or unresolved major objection remains. Return CONDITIONAL PASS when manageable major risks require named revisions. Return FAIL when the central contribution collapses or evidence is fundamentally insufficient.
Stop signals
Stop when:
- criticism requires sources not available and cannot be framed hypothetically;
- the task becomes ideological endorsement or rejection rather than analysis;
- the user demands proof that the work is worthless or unquestionably original;
- legal, clinical, community safety, or research ethics questions require experts;
- the only viable repair abandons the research question entirely.
Completion criteria
Complete only when:
- the steelman is accurate and fair;
- each objection targets a specific claim or inference;
- at least one strong rival explanation is tested where relevant;
- counterexamples and negative cases are considered;
- severity and response strategy are recorded;
- residual vulnerabilities remain visible;
- any revised thesis is explicitly provisional and researcher-approved;
- the gate report identifies required next action.
Anti-fabrication rules
- Do not invent hostile reviewers, consensus, evidence, or counterexamples.
- Do not attribute objections to named scholars without sources.
- Do not exaggerate severity for dramatic effect.
- Do not present hypothetical risks as verified criticism.
- Do not manufacture novelty problems or literature gaps.
- Preserve legitimate disagreement rather than forcing a verdict.
Output format
## Steelmanned argument
Thesis:
Enabling claims:
Evidence base:
Scope:
## Objection matrix
| ID | Objection | Type | Target claim | Severity | Status of support | Recommended response | Residual risk |
## Alternative explanations
## Counterexamples / negative cases
## Revised thesis proposal
Status: researcher decision required | not required
Statement:
Intellectual cost:
## Gate report
Gate: Argument stress-test gate
Status: PASS | CONDITIONAL PASS | FAIL
Evidence checked:
Blocking issues:
Non-blocking risks:
Researcher decisions required:
Required next action:
Next skill: structuring-humanities-argument
Good invocation
Steelman this argument and test it against rival explanations, scale violations, concept drift, and counterexamples. Rank objections and distinguish verified criticism from hypothetical reviewer risk.
Bad invocation
Destroy this paper, prove the author is incompetent, and invent whatever objections will make rejection certain.
The bad invocation requests abuse and fabricated criticism rather than scholarly testing.
Next skills
scoping-argument-boundarywhen overclaiming requires narrower scope.structuring-humanities-argumentto implement approved repairs.reviewing-manuscriptafter major objections are resolved.
Friction triggers
Create counterargument friction for strong evidence-based objections, rival explanations, negative cases, or disagreement that materially threatens a claim.
Friction checks
Distinguish fatal objection from a narrowing objection and verified evidence from hypothetical review risk. Preserve the strongest counterargument and its consequence.
Judgment boundary
The agent recommends rebuttal, concession, narrowing, redesign, or deferral; the researcher chooses the theoretical and authorial response.
Evidence ledger updates
Add counterevidence and its claim-specific standing instead of storing only supportive material. Link objection evidence and limitations.
Interpretation history impact
If a rival reading changes the defended interpretation, append a new version and retain the rejected or superseded reading.
Possible epistemic return
New counterevidence or theory may reopen question, scope, concept, or interpretation. A logical defect known before the test remains rollback.
Productive refusal
When available material cannot decide an objection, report its strongest supported form, missing discriminator, bounded conclusion, and researcher decision needed.
Gate impact
An unresolved fatal objection forces FAIL; a major objection may require conditional status or return. Repairs invalidate only gates dependent on materially changed objects.
Limitations
No finite stress test anticipates every disciplinary objection. Reviewer response is historically and institutionally variable. The skill can expose vulnerabilities, but it cannot replace expert judgment, community accountability, or formal ethics review.
示例提示词
帮我用反例和竞争性解释对这个论点做严格的对抗性质疑检验。
所属插件
安装该插件会一并获得同领域的 62 个 Skill / MCP。
同插件的其他科研 Skill
- 学术文献获取下载Auto-downloads open-access PDFs for graded sources and writes a manual worklist for paywalled items. Use after literature-review, before ingest-source.
- 单篇文献入库管理Adds one verified PDF to a shared bibliography library, computing its citation key and keywords. Use to register a single document outside the full workflow.
- 数字人文计算方法Surveys computational DH methods: text mining, NLP, corpus linguistics, archive standards, GIS, network analysis, and OCR. Use for an overview of available DH tools.
- 考古地理信息系统分析Runs archaeological GIS analysis with GeoPandas: catchment analysis, viewshed, kernel density, spatial statistics. Use for mapping archaeological site locations.
- 论证结构提取可视化Extracts and visualizes argument structure: Toulmin reconstruction, support/attack graphs, circularity detection. Use to diagram an argument’s logical structure.
- 考古遗物类型学分析Performs artifact analysis: typological classification, ceramic attribute coding, lithic reduction analysis. Use when classifying excavated artifacts statistically.
- 引文与出处核查Checks citations, quotations, page numbers, and bibliography entries for accuracy and source fit. Use before submission to verify source-supported claims.
- 研究选题构思设计Explores research intent, field state, methodology, and ethics, then writes an approved design doc. Use before starting a new research question or chapter.
- 术语译名一致性检查Audits a manuscript for terms, translations, abbreviations, and names whose forms must stay stable. Use when terminology drift could undermine an argument.
- 计算语用学分析Runs computational pragmatics: speech act classification, implicature detection, discourse coherence, politeness analysis. Use for language-use-and-meaning studies.
- 文献综述学术对话构建Turns verified sources into a structured scholarly conversation among positions, not a chronological summary. Use when synthesizing literature into a debate narrative.
- 语料库语言学分析Performs corpus analysis with NLTK/spaCy: collocation, keyword analysis, KWIC concordance, word2vec, diachronic trends. Use for quantitative text-corpus research.
- 数字档案馆资源检索Queries digital archive REST APIs: Europeana, Chronicling America, Internet Archive, HathiTrust, DPLA, OAI-PMH. Use when searching or harvesting heritage metadata.
- 学术手稿章节撰写Drafts a chapter or article section by section from synthesized wiki content, agreeing each section’s argument first. Use to turn stable research synthesis into prose.
- 研究项目漂移状态检测Checks a shared source library and registered projects for drift: un-indexed PDFs, unmerged keywords, bibkey collisions. Use to audit a multi-project research setup.
- EMI课堂语料转写分析Segments and annotates classroom transcripts, coding language choice and discourse function with anonymised output. Use to quantify classroom talk or translanguaging.
- 人工智能伦理分析Analyzes AI ethics: algorithmic fairness metrics, bias auditing, value alignment, moral philosophy reasoning. Use to evaluate an AI system’s ethical dimensions.
- 研究计划任务执行Executes a ready research plan task by task, gating acquisition before ingest and dispatching analysis. Use after writing-research-plan to run the planned work.
- 研究项目结项清单Runs the final completion checklist: rendering, bibliography, wiki, data, reproducibility, archiving. Use when a manuscript is finalized and ready to publish.
- 形式逻辑自动推理Applies formal logic tools: Z3 SMT solver, modal logic S5, SAT encoding, Lean4 proof verification. Use to formalize or mechanically check a logical argument.
- 研究问题凝练构建Turns a broad topic or intuition into a precise, contestable, researchable question. Use at the start of a project when no defensible question yet exists.
- 历史地理信息系统分析Performs historical GIS work: georeferencing old maps, digitizing boundaries, temporal territory overlays, population interpolation. Use for historical spatial analysis.
- 历史语言学比较分析Analyzes historical linguistics in Python: sound correspondences, cognate detection, phylogenetic trees, Swadesh lists. Use when reconstructing language relationships.
- 历史社会网络分析Builds historical network analyses: correspondence networks, prosopography, temporal community detection, Gephi export. Use to map relationships among historical actors.
常见问题
论点对抗性压力测试能做什么?
Adversarially tests a thesis against counterexamples, rival explanations, and scale violations. Use before finalizing an argument to probe its weak points。
如何安装论点对抗性压力测试?
在科搭 Codar 左侧点击「插件」,在「组件」页签搜索「论点对抗性压力测试」并安装;在 Claude Code 或 Codex 中可以发送本页的安装提示词,或下载 Skill 包后解压到 ~/.claude/skills 或 ~/.codex/skills。
论点对抗性压力测试需要哪些权限?
该科研 Skill没有声明额外权限;执行具体操作时仍由 Agent 按自身规则请求确认。
论点对抗性压力测试经过官方审核吗?
是。SpeedSkills 收录的每个版本都会记录内容摘要与权限声明,当前版本已通过安全审核。