evidence-gateGenerates evidence obligations for a claim or action, evaluates existing evidence against them, and returns a structured verdict (PASS / SOFT_PASS / BLOCK /...
Install via ClawdBot CLI:
clawdbot install shanicky/evidence-gateGrade Fair — based on market validation, documentation quality, package completeness, maintenance status, and authenticity signals.
Calls external URL not in known-safe list
https://json-schema.org/draft/2020-12/schemaAudited Apr 17, 2026 · audit v1.0
Generated Mar 22, 2026
When a DevOps engineer identifies a potential root cause like 'database connection pool exhaustion' during an outage, this skill can gate the claim by requiring evidence such as monitoring dashboards showing connection spikes, application logs with timeout errors, and comparison with baseline metrics. It ensures the diagnosis is backed by multiple data sources before recommending a restart or scaling action.
Before approving a production configuration change like 'increasing server timeout limits', the skill evaluates evidence obligations including performance test results, historical incident reports related to timeouts, and peer review approvals. It helps prevent unsafe deployments by flagging missing evidence like A/B test outcomes.
When an AI agent suggests 'deleting old user logs for GDPR compliance', the skill gates this by requiring evidence such as data retention policy documents, audit trail logs confirming data age, and legal team sign-offs. It ensures high-risk actions are justified to avoid regulatory penalties.
For a medical AI proposing 'the patient's symptoms indicate condition X', the skill requires evidence like lab test results, imaging reports, and differential diagnosis analysis. It blocks premature conclusions by highlighting conflicting evidence from patient history, promoting safer clinical decision support.
Before an AI recommends 'investing in a volatile stock is safe', the skill evaluates evidence obligations including market trend analyses, risk assessment models, and diversification plans. It downgrades the claim if only historical price data is available, guiding users to gather volatility metrics and expert consultations.
Offer this skill as part of a platform subscription for incident management tools, charging per active user monthly. It integrates with existing monitoring systems to provide auditable evidence gates, reducing mean time to resolution and preventing costly misdiagnoses.
Sell enterprise licenses to organizations in regulated industries like finance or healthcare, where evidence gating is critical for compliance. Revenue comes from one-time setup fees and annual maintenance contracts, with customization for specific regulatory frameworks.
Provide a free tier for individual developers or small teams with basic evidence gating, and premium features like advanced analytics and integrations for larger organizations. Monetize through upsells to paid plans that include priority support and enhanced security features.
💬 Integration Tip
Integrate this skill by hooking it into existing decision-making workflows at key conclusion points, ensuring it runs selectively to avoid overhead while providing structured evidence checks.
Scored Apr 19, 2026
Multi-source deep research agent. Searches the web, synthesizes findings, and delivers cited reports. No API keys required.
Transparent, rigorous research with full methodology — not a black-box API wrapper. Conducts exhaustive investigation through mandated 2-cycle research per t...
Seven-stage deep research pipeline for substantive research questions. Activates when user asks for deep research, 深度研究, research report, evidence-backed inv...
Perform deep research using Claude by autonomously gathering, analyzing, and synthesizing detailed information across diverse topics on request.
全生命周期质量保障专家,覆盖需求、设计、编码、测试、上线、运维全阶段的质量保障活动。 融合缺陷预防(逆向操作、依赖踏空、并发冲突、新旧兼容、状态迁移、因果判定)与质量度量、持续改进三大支柱, 帮助团队建立「预防-评审-度量-改进」的质量闭环,实现软件质量的持续提升。 Use when: - 需求/设计/编码/测试...
专家团自动组建技能(反面教材)。核心观点:AI只认事不认人,没必要模拟人类多角色协作方式。本技能演示为什么不需要这样做——AI应该直接面对任务,而不是模拟人类团队。触发词:组建专家团、专家协作、团队完成任务、自动组建团队、expert team、专家团、团队协作完成任务。