Scientific Critical Thinking
K-DenseEvaluates scientific claims and evidence quality: experimental validity, biases, confounders and evidence grading.
By agent · Hermes
Skills in the open SKILL.md format that Hermes reads. Each one has a plain-language summary and the command to install it.
$ npx skills add <author>/<skill> --agent hermes-agentReplace the placeholder with the author and skill name from the skill's page. Start a new Hermes session afterwards so the skill is picked up.
Evaluates scientific claims and evidence quality: experimental validity, biases, confounders and evidence grading.
Builds slide decks for research talks, conference presentations and thesis defences.
Creates and audits publication-ready scientific figures with Matplotlib, Seaborn or Plotly.
Drafts, revises and audits scientific manuscripts with evidence provenance and reporting-guideline coverage.
Machine learning in Python with scikit-learn: classification, regression, clustering, model evaluation and tuning.
Statistical visualisation with Seaborn: distributions, relationships and categorical comparisons.
Guides statistical analysis of research data: test selection, assumption checks, effect sizes and APA-formatted reporting.
Statistical modelling with statsmodels: OLS, GLM, mixed models and ARIMA, with diagnostics and inference.
Runs agent-browser on AWS Bedrock AgentCore cloud browsers, using AWS credentials.
The core usage guide for agent-browser: the snapshot-and-ref workflow, navigation, element interaction and data extraction.
Reverse-engineers a website's internal API from recorded browser traffic and generates a standalone client for it.
Explores and tests a web application systematically to find bugs and UX issues.
Automates Electron desktop apps such as VS Code, Slack and Notion with agent-browser.
Accesses and tests Vercel deployments that sit behind Vercel Authentication, SSO or Deployment Protection.
Works in Slack through browser automation: check unread channels, send messages and search conversations.
Runs agent-browser and Chrome inside Vercel Sandbox microVMs for browser automation from a Vercel app.
React composition patterns such as compound components, for refactoring components with too many boolean props.
Deploys applications and websites to Vercel, including preview deployments.
React and Next.js performance guidelines from Vercel Engineering, for writing and reviewing code.
React Native and Expo practices for performant mobile apps: lists, animations and native modules.
Guides implementing animations with React's View Transition API.
Deploys and manages Vercel projects with token-based authentication instead of interactive login.
Finds cost and performance improvements for deployed Vercel projects from metrics, usage and code scans.
Reviews UI code against Vercel's Web Interface Guidelines, including accessibility and UX.
Questions
Often. A skill is a folder of instructions and scripts in the open SKILL.md format, which Hermes reads. A skill that depends on tools only one agent provides will not carry over, so read the skill's source before you rely on it.
Each skill page has an install command. Run it in a terminal, then start a new Hermes session so the skill is picked up.
An instructions file is loaded in every session and describes the project. A skill is loaded only when a task calls for it, so it can be long and specific without costing context on every turn.