Designing for AI Agents: McP—Jessica Kerr  🔗 **Source:** [Designing for AI Agents: McP—Jessica Kerr](https://www.youtube.com/watch?v=yop6v0KxRwM) > [!note]+ Document Notes > tags: [MCP, AI Agents, Product Design] > author: [Jessica Kerr] > source: [YouTube] > reference: [https://www.youtube.com/watch?v=yop6v0KxRwM] > date created: {{date:YYYY-MM-DD}} {{time:HH:mm}} > > # 🏷️ Title: Designing for AI Agents: MCP—Jessica Kerr > <!-- deep-read:sections start --> ## 🧭 Table of Contents - [📝 Summary](#-summary) - [📖 Main Content](#-main-content) - [🗒️ Notes](#️-notes) - [📚 References](#-references) - [🗣️ Paraphrase](#️-paraphrase) - [📰 Newsletter](#-newsletter) - [✅ Action Items](#-action-items) - [🧪 Hypotheses](#-hypotheses) - [📊 Methodology](#-methodology) - [❓ Questions](#-questions) - [🔮 Future Research](#-future-research) - [💡 Implications](#-implications) - [🧠 Learning Styles Assessment](#-learning-styles-assessment) - [📊 Accessibility Assessment](#-accessibility-assessment) - [🌍 DEIB Assessment](#-deib-assessment) - [✅ Logical Validity](#-logical-validity) - [🛡️ DARVO Analysis](#-darvo-analysis) - [🎯 Conclusion](#-conclusion) ### 🔗 References - Video: Designing for AI Agents: MCP—Jessica Kerr. https://www.youtube.com/watch?v=yop6v0KxRwM - Honeycomb observability platform. https://www.honeycomb.io/ - Anthropic Model Context Protocol overview. https://modelcontextprotocol.io/ *(If inaccessible, use the YouTube context above.)* - Playwright testing framework. https://playwright.dev/ - Figma design platform. https://www.figma.com/ ## 📝 Summary - Jessica Kerr explains MCP’s role in AI agent workflows. [Video](https://www.youtube.com/watch?v=yop6v0KxRwM) - MCP connects agents to live context securely and flexibly. - Honeycomb uses MCP to bring production data to daily tools. - Stakeholders include designers, developers, sales, and data teams. - Risks include token bloat, weak error design, and tool sprawl. - Bottom line: Treat MCP as a third interface with living documentation. ## 📖 Main Content - MCP enables agents to fetch context from services and tools. - An MCP server exposes prompts and tools to the agent client. - The agent mediates auth, tool calls, and model interactions. [Video](https://www.youtube.com/watch?v=yop6v0KxRwM) - Fewer, clearer tools reduce tokens and improve reliability. - Opinionated responses guide models toward useful next steps. - Honeycomb teams saw better feedback loops across code and ops. [Honeycomb](https://www.honeycomb.io/) - Formats like ASCII tables can beat JSON for brevity and clarity. - Uncertainties remain around cross-model consistency and security policies. - Good evaluations track task success, time, calls, and token spend. ### 🗒️ Notes #### ⏱️ [00:06](https://www.youtube.com/watch?v=yop6v0KxRwM&t=6) Design excitement - Amanda Bernard values AI’s design impact. - Honeycomb needs clear UI for complex telemetry. #### ⏱️ [02:19](https://www.youtube.com/watch?v=yop6v0KxRwM&t=139) What MCP connects - MCP links agents, models, users, and tools. - Calendar, tasks, and Figma can integrate. #### ⏱️ [04:27](https://www.youtube.com/watch?v=yop6v0KxRwM&t=267) Not a classic API - LLMs accept changing text outputs. - MCP allows flexible, evolving responses. #### ⏱️ [06:41](https://www.youtube.com/watch?v=yop6v0KxRwM&t=401) Living documentation - Client authenticates and requests tools. - Server lists tools with instructions. #### ⏱️ [08:40](https://www.youtube.com/watch?v=yop6v0KxRwM&t=520) Minimize tools - Too many tools waste tokens. - Disable rarely used endpoints. #### ⏱️ [10:43](https://www.youtube.com/watch?v=yop6v0KxRwM&t=643) Opinionated outputs - Provide helpful guidance in responses. - Use compact formats like ASCII. #### ⏱️ [12:42](https://www.youtube.com/watch?v=yop6v0KxRwM&t=762) Feedback loops - Improve error messages with expertise. - Let production signals refine design. #### ⏱️ [14:37](https://www.youtube.com/watch?v=yop6v0KxRwM&t=877) Evaluation practice - Score success, latency, calls, and tokens. - Run nightly benchmarks across agents and models. ## 📚 References - [Designing for AI Agents: MCP—Jessica Kerr](https://www.youtube.com/watch?v=yop6v0KxRwM) - [Model Context Protocol](https://modelcontextprotocol.io/) - [Honeycomb](https://www.honeycomb.io/) - [Playwright](https://playwright.dev/) - [Figma](https://www.figma.com/) ## 🗣️ Paraphrase - MCP gives agents secure access to up-to-date context from many systems. - Teams should trim tool lists, craft guidance, and measure outcomes continuously. ## 📰 Newsletter - MCP is reshaping agent UX by centering real context, not screens. Watch Jessica Kerr’s talk, then evaluate your agent tool lists. Start small, write opinionated responses, and measure token costs. Watch: https://www.youtube.com/watch?v=yop6v0KxRwM • Docs: https://modelcontextprotocol.io/ • Honeycomb: https://www.honeycomb.io/ ## ✅ Action Items - DevRel - Record a 6-minute MCP primer for OSS maintainers. - Demo “fewer tools” vs “API dump” with metrics. - Developer advocacy: plans, talks, demos, and measurable outcomes. - Publish MCP design checklists with token budgets and examples. - Education: workshops, curricula, office hours, mentorship programs. - Create a lab: secure OAuth flow for an MCP server. - Developer experience: docs, SDKs, examples, templates, quickstarts. - Ship a minimal MCP starter with tests and tracing hooks. - Developer marketing: campaigns, content, funnels, and attribution. - Run a “Context over UI” campaign with case studies. - Community management: moderation, code of conduct ops, surveys. - Host a roundtable on MCP evaluation metrics. - Internal/external/cross-project: align goals, SLAs, and feedback loops. - Define SLAs for MCP error quality and latency. - OSPO - Policy & Governance: contribution policies, decision rights, charters. - Add MCP service contribution and review steps. - License Compliance: policies, scanning, review, and training cadence. - Audit MCP servers for third-party license obligations. - Security Management: vuln intake, embargo policy, signing, attestations. - Require signed server builds and rotated credentials. - Software Inventory: SBOM strategy, ownership, freshness SLOs. - Track MCP servers and dependencies in SBOMs. - Developer Support: training, tooling, legal guardrails, consultation hours. - Offer quarterly MCP security clinics. - Community Engagement: upstream funding, MoUs, operator agreements. - Sponsor reference MCP servers for common tools. - Strategic Alignment: KPIs, OKRs, executive reporting, initiative roadmaps. - Report agent success rates and token costs quarterly. - Community - Maintainers: roadmap clinics, governance templates, release hygiene. - Add MCP error-message standards to templates. - Contributors: onboarding paths, “first issue” rotations, pairing sessions. - Label “good first MCP server” issues. - Sponsors: transparent budgets, milestones, deliverables, impact reports. - Fund evaluation harnesses and datasets. - Ecosystem: cross-project incident drills and coordinated disclosures. - Practice MCP credential leak scenarios. - WordPress - Core: priorities, issue triage, performance budgets, release criteria. - Explore MCP-like adapters for .org infrastructure. - Core Performance: profiling plans, metrics dashboards, perf bounties. - Track agent-added script costs in wp-admin. - Core AI: responsible AI guardrails, evaluation sets, pilot projects. - Pilot agents for triage with strict privacy. - Core Program: coordination rituals, cross-team dependencies, retrospectives. - Share evaluation dashboards with teams. - Mobile: app backlog, telemetry review, offline UX improvements. - Evaluate agent tasks on devices with spotty networks. - Playground: browser labs, reproducible bugs, education modules. - Build an MCP sandbox for tutorials. - Community: meetup kits, WordCamp playbooks, speaker mentorship. - Provide an “Agents and MCP” talk kit. - Training: lesson plans, learning paths, certification pilots. - Create MCP server certification basics. - Marketing: campaigns, narratives, social editorial calendars. - Highlight accessible, low-token agent designs. - Support: response SLAs, knowledge base gaps, escalation paths. - Document common agent failure modes and fixes. - Design: UX audits, design tokens, research backlogs. - Define patterns for agent messages and errors. - Documentation: DevHub/HelpHub gaps, examples-first patterns, style guide. - Add checklists for agent tool definitions. - Photos: content drives, moderation guidelines, taxonomy hygiene. - Curate diagrams explaining agent flows. - TV: submission guides, caption quality, content indexing. - Encourage agent tutorial subtitles and chapters. - Polyglots: translation sprints, glossary alignment, locale coverage. - Localize agent error messages. - Accessibility: WCAG reviews, tooling, issues heatmap, audits cadence. - Audit agent UIs for keyboard and SR parity. - Meta: .org roadmap, API reliability, contributor tools. - Offer MCP-like APIs for community tools. - CLI: command coverage, UX docs, plugin scaffolds, smoke tests. - Scaffold MCP-ready plugins and health checks. - Hosting: cross-host checklists, defaults, telemetry for reliability. - Monitor agent traffic and rate limiting. - Tide: scoring transparency, false-positive review loops, plugin outreach. - Consider agent and MCP quality signals. - Openverse: ingestion pipelines, moderation policy, education tie-ins. - Provide icons for agent states. - Themes: review throughput, lints, mentoring, test matrices. - Review agent UI components for accessibility. - Plugins: review throughput, security liaisons, provenance checks. - Require provenance for agent tool calls. - The WPCC - Programs: fellowship tracks, microgrants, maintainer stipends. - Fund OSS MCP servers for community tools. - Projects: governance clinics, documentation sprints, release teams. - Publish MCP design and evaluation playbooks. - Funding experiments: matched grants, milestone payouts, transparent ledgers. - Sponsor token-budget benchmarks with open data. - FAIR - Technical independence: local metadata, provenance, and signature checks. - Sign MCP artifacts and verify provenance. - Policies and covenants: funding transparency, sunset clauses, anti-capture. - Disclose sponsor influence on server design. - Productized governance artifacts: charters, operator agreements, scorecards. - Release MCP server scorecards with KPIs. ## 🧪 Hypotheses - Reducing tools by 50% improves task success by 10%. - Opinionated responses cut tool calls per task by 20%. - ASCII tabular outputs reduce token use by 15%. - Nightly evaluations improve weekly success rates by 8%. - Better error messages reduce repeated failures by 25%. ### 📊 Methodology - A/B test tool lists across matched tasks and users. - Compare JSON, CSV, and ASCII formats on tokens and accuracy. - Run nightly evaluation suites across agents and models. - Instrument success, latency, token spend, and retries. - Review error logs with rubric-based ratings. ### ❓ Questions - Which guidance patterns generalize across leading models? - How should we budget tokens per request reliably? - What auth patterns best prevent credential leaks? - When should humans review high-risk agent actions? ### 🔮 Future Research - Standard metrics for agent evaluation across ecosystems. - Safer action execution with staged confirmations. - Schema hints that travel well across models. - Shared open datasets for MCP benchmarking. ### 💡 Implications - Practice: Design MCP as a primary interface, not glue. - Policy: Establish evaluation and error quality standards. - Security: Rotate credentials and sign builds. - Education: Teach opinionated response design. - Funding: Back open MCP servers and benchmarks through The WPCC. ### 🧠 Learning Styles Assessment - Reading level estimate: Flesch–Kincaid Grade ≈ 10. - Visual learners - Present: diagrams and flows implied in talk. - Improve: add architecture and state charts. - Aural learners - Present: narrated examples and rationale. - Improve: 3-minute audio summary. - Verbal learners - Present: definitions and comparisons. - Improve: glossary and pattern names. - Kinesthetic learners - Present: prompts and tool steps. - Improve: reproducible labs and checklists. - Logical learners - Present: evaluation metrics and loops. - Improve: decision trees and KPIs. - Social learners - Present: team feedback loops. - Improve: cohort reviews and forums. - Solitary learners - Present: self-paced viewing. - Improve: reflection prompts and quizzes. - Naturalistic learners - Present: limited connections. - Improve: sustainability of token usage. ### 📊 Accessibility Assessment - Screen reader semantics - Present: clear structure in examples. - Improve: explicit ARIA guidance for agent UIs. - Keyboard navigation - Present: emphasis on focus and flow. - Improve: escape behavior and traps testing. - Contrast and readability - Present: not evaluated. - Improve: ensure 4.5:1 contrast in agent UIs. - Text resizing and reflow - Present: not evaluated. - Improve: verify 200% zoom with no clipping. - Media and captions - Present: YouTube captions available. - Improve: downloadable transcript and audio description. - Forms and interactions - Present: tool call patterns shown. - Improve: larger targets and inline validation. - Standards reference - Map open gaps to WCAG 2.2 AA criteria. ### 🌍 DEIB Assessment - Cultural breadth - Present: industry-neutral examples. - Improve: global case studies and translations. - Inclusive language - Present: neutral, clear terms. - Improve: plain-language summaries for newcomers. - Neurodiversity support - Present: chunked concepts and repetition. - Improve: icon+text cues and predictable layouts. - Representation - Present: varied roles mentioned. - Improve: inclusive imagery in examples. - Psychological safety - Present: supportive tone and learning focus. - Improve: published community guidelines for agent usage. ### ✅ Logical Validity - Strong reasoning - Causal: fewer tools reduce tokens and confusion. - Analogical: “third interface” clarifies differences from APIs and UIs. - Fallacies - Possible hasty generalization on ASCII benefits. - Fix: validate across models and tasks. - Tone of voice - Practical, confident, and instructional. ### 🛡️ DARVO Analysis - Deny - No denial or minimization tactics observed. - Attack - No attacks on credibility observed. - Reverse victim/offender - Not applicable to this instructional talk. ## 🎯 Conclusion - MCP reframes agent design around context, guidance, and evaluation. - Start lean, guide the model, measure outcomes, and refine with feedback loops. <!-- deep-read:sections end --> [[3. Resources/Readwise/Full Document Contents/Articles/Designing for AI Agents McP—Jessica Kerr|Full Document Contents]] ## Highlights - manda is excited about MCP because it brings that sweet production data right to where people are. And that's why MCP is an MVP. I'm illustrating this with our internal experiences of implementing an MCP very recently so that you can learn what we've learned. For instance, internally, our salespeople could log into Honeycomb to find out what people are doing in the product, but they don't. However, they will type in Claude, Desktop. ([View Highlight](https://read.readwise.io/read/01k6dqzvwvd9925ftv9z7ffp62))