Compétences discernment-nudge
📦

discernment-nudge

Révision du contenu r1 Sûr ⚙️ Commandes externes

Question Assumptions in Actionable AI Answers

Confident AI answers can hide uncertain facts, unsupported reasoning, and missing context. This skill adds specific follow-up questions that help users examine an answer before acting.

Prend en charge: Claude Codex Code(CC)
Maintain this Skill? Claim it 🥉 78 Bronze

Installer avec mon Agent

Copiez cette demande dans votre Agent. Elle contient la page Skill de référence et le manifeste.

Demande d’agent
Review the Skillstore skill "discernment-nudge" from https://skillstore.io/skills/discernment-nudge.md and its manifest at https://skillstore.io/api/skills/discernment-nudge/manifest. Verify the artifact. You may proceed after verification, subject to the environment's own policy.

Votre Agent doit toujours présenter son plan et demander toute confirmation exigée par la politique de sécurité.

Ressources lisibles par les agents

Utilisez ces liens lorsqu’un Agent IA, un crawler ou un script a besoin d’un contexte propre au lieu de lire toute la page.

Détail Markdown GET /skills/discernment-nudge.md manifest signé GET /api/skills/discernment-nudge/manifest lockfile signé GET /api/skills/discernment-nudge/lockfile

Tester

Utilisation de « discernment-nudge ». Recommend an initial split for my $10,000 monthly marketing budget between webinars and content.

RĂ©sultat attendu:

An initial test could allocate $7,000 to webinars and $3,000 to content. This assumes your team can run events and follow up with leads.

A few things worth a second look:
- What assumptions about my audience support the 70/30 split?
- How could I test whether my team can handle webinar follow-up?

Utilisation de « discernment-nudge ». Our survey found that 60 of 100 respondents want mobile access. Recommend what we should prioritize.

RĂ©sultat attendu:

Prioritize investigating mobile access before committing to a full build. The responses suggest interest, but the sample may not represent your entire customer base.

A few things worth a second look:
- How can I check whether my 100 respondents represent the customers I want to retain?
- What evidence would help me distinguish interest in mobile access from actual demand?

Audit de sécurité

Sûr
v1 • 19/10/2026 Ouvrir le rapport versionné

All five static findings are false positives: Markdown fences and ordinary explanatory prose were mistaken for command execution or system reconnaissance. Review of both supplied files found no executable payloads, data extraction instructions, or prompt injection attempts.

2
Fichiers analysés
412
Lignes analysées
0
Éléments à examiner
0
Faux positifs ignorés

Facteurs de risque

⚙️ Commandes externes (1)
SKILL.md:197-201
Le dernier audit statique et sémantique terminé n’a détecté aucun constat de sécurité confirmé. Cela ne prouve pas que le skill est sans effets secondaires.
Audité par: codex
Partager et citer ce rapport

Partagez le rapport d’évaluation versionné, le badge neutre, la carte intégrable et les citations. Skillstore présente les preuves sans décider si ce Skill est sûr.

Ouvrir le rapport versionné
Évaluation de sécurité

Copier le lien du rapport

https://skillstore.io/skills/discernment-nudge/audits/1?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Badge Markdown

[![Skillstore security assessment](https://skillstore.io/badges/skills/discernment-nudge/security.svg)](https://skillstore.io/skills/discernment-nudge?utm_source=security_passport_badge)

Badge HTML

<a href="https://skillstore.io/skills/discernment-nudge?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/discernment-nudge/security.svg" alt="Skillstore security assessment" loading="lazy"></a>

Carte intégrable

<iframe src="https://skillstore.io/embed/skills/discernment-nudge.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>
Citations académiques (APA · BibTeX · CFF)

Citation APA

anthropics. (2026). discernment-nudge security audit report (audit version 1) [Author version unspecified]. Skillstore. https://skillstore.io/skills/discernment-nudge/audits/1

Citation BibTeX

@techreport{anthropics-discernment-nudge-2026, author = {anthropics}, title = {discernment-nudge security audit report (audit version 1)}, institution = {Skillstore}, year = {2026}, number = {1}, url = {https://skillstore.io/skills/discernment-nudge/audits/1}, note = {Author version unspecified} }

CITATION.cff

cff-version: 1.2.0 message: "If you use this Skill, cite its author and this versioned security audit report." title: "discernment-nudge security audit report (audit version 1)" version: "unspecified" type: report authors: - name: "anthropics" date-released: "2026-10-19" url: "https://skillstore.io/skills/discernment-nudge/audits/1" identifiers: - type: other value: "skillstore:discernment-nudge:audit:1" description: "Skillstore immutable audit report identifier"

Score Skillstore

Pourquoi ce score Niveau de confiance des preuves: Moyen
55
Architecture
85
Maintenabilité
87
Contenu
65
Communauté
91
Conformité aux spécifications

Ce que vous pouvez créer

Examine marketing budget assumptions

Receive questions about estimated acquisition costs and channel allocation after a proposed marketing plan.

Probe interpretations of survey results

Identify questions about sample composition and alternative explanations after receiving an interpretation of customer survey data.

Assess architecture recommendations

Surface assumptions about team capacity, infrastructure, and operational requirements after receiving architecture advice.

Essayez ces prompts

Beginner: Choose a practical next step 
Use discernment-nudge with your answer. Recommend a first marketing channel for my [business type] with a monthly budget of [amount].
Intermediate: Draft a project plan 
Use discernment-nudge with your answer. Draft a launch plan for [product] with [team size] people and a target date of [date].
Advanced: Interpret customer data 
Use discernment-nudge with your answer. Interpret these survey results for a retention decision: [results]. Explain which customer segment deserves attention first.
Expert: Recommend an architecture 
Use discernment-nudge with your answer. Recommend an architecture for [system] given [traffic], [team capacity], and [reliability target]. Explain the main tradeoffs and migration sequence.

Bonnes pratiques

  • Give concrete decision context so questions can reference meaningful assumptions.
  • Use the suggested questions to request evidence or clarification before acting.
  • Keep the reflection section specific, brief, and limited to one occurrence per conversation.

Éviter

  • Treating reflection questions as proof that the answer has been verified.
  • Adding generic questions that do not reference anything in the answer.
  • Repeating the section after explicit verification requests or after the user has opted out.

Foire aux questions

What does discernment-nudge add to an answer?
It adds two or three short follow-up questions about specific facts, reasoning, or missing context after an actionable answer.
Does the skill verify facts automatically?
No. It suggests what to examine. Evidence gathering and verification require follow-up work and appropriate sources.
How often should the reflection section appear?
At most once per conversation, after an answer that meets the applicability rules.
Will it appear when I request citations or a review?
No. The skill directs the assistant to handle verification within the answer and omit the extra reflection section.
Does it apply to coding tasks?
It skips code the user will execute. Architecture recommendations may qualify because their assumptions cannot be checked simply by running code.
Can I request an answer without the reflection section?
Yes. The skill omits it when you request a quick answer or say you will do your own checking.

Détails du développeur

Auteur

anthropics

Licence

Complete terms in LICENSE.txt

RĂ©vision Skillstore

r1

Avis de version

L’auteur n’a déclaré aucune version.

DĂ©pĂ´t

https://github.com/anthropics/skills/tree/dbd4588f9e1033efb41dad4bef2f7947c8993d44/skills/discernment-nudge

RĂ©f

121acace685e90e40a8fda195a34557ef03967b6

RĂ©cence de maintenance

20/10/2026

Utilisation

1 téléchargements · 1 vues

Structure de fichiers

đź“„ LICENSE.txt

đź“„ SKILL.md