The New Product Surface Is Where the Work Turns
The next AI product advantage may not be a better answer. It may be better placement: carrying customer and domain context into the moment where work turns into action.
The Board’s AI Story Has a Half-Life
A frontier model can move from launch to unavailable in a week. Boards need more than AI education. They need a disciplined way to recognize when a changing capability invalidates the story guiding products, promises, and investment.
The Agent Is Not the Moat
The agent gets the product name and the animated workflow. The durable advantage is accumulating underneath it: distinctive signal, legitimate customer access, and a learning loop that makes the next response more relevant.
Letaido Puts the Marketing Dashboard Inside the Loop
Ahrefs' Letaido points to a shift in marketing software: the dashboard is becoming a window into a persistent agent loop. Routine sensing and analysis can run continuously, while human attention moves toward objectives, interpretation, customer meaning, and creative direction.
The Missing Product in AI-Assisted Publishing
AI-assisted publishing does not need a perfect detector first. It needs a better byline: a concise, layered signal that explains what AI did, what human judgment remained, and who stands behind the finished work.
When AI Makes Polish Cheap, the Making Becomes the Marketing
Generative AI is making competent polish abundant. A surprise Chinese animation hit suggests the next creative advantage may belong to brands that make their choices, constraints, and authorship visible.
Can a SaaS Output Carry Its Own Renewal Case?
Google's AI ad disclosure points to a broader product shift: when agents become the user, SaaS products may need outputs that carry their own commercial context into budget and renewal decisions.
The AI-Readable Ad Shelf Is Open. Who Is Buying?
AI-readable advertising moves the ad shelf upstream. Brands can now buy presence in publisher pages that AI systems may use to form product recommendations, but a paid placement is still not a recommendation.
An AI Label Can Tell the Truth and Still Reduce Trust
The best AI disclosure is not always the longest. Audiences judge whether AI strengthened the offer or removed the expertise, care, identity, or accountability they thought they were buying.
Who Owns the Customer Context When Agents Can Orchestrate the Campaign?
As OpenAI and Google build delegated marketing work surfaces, campaign execution is becoming easier to rent. The scarce asset is the customer context that tells those systems what should happen next.
The Unit of Creation Is No Longer the Asset. It's the Operating Loop.
Adobe’s June 2026 expansion is not “another creative AI feature drop.” It is a product shift that relocates the surrounding workflow into the suite itself. When creative and marketing platforms can execute outcome specs with editable artifacts and clear escalation boundaries, the unit of creation b…
The New Scoreboard Is Placement: What Google's Gemini Sequence Actually Reveals
Reports say Google’s best coding model is about six months behind the frontier. Yet Google shipped Gemini into Docs, Sheets, Slides, Drive, and Android surfaces. The lag is the decoy. The move is the point: competitive advantage is rotating from model IQ to placement economics and task assignment,…
The Taste Contract: Adobe's New Coworker Turns Brand Judgment Into a Product Primitive
Adobe’s CX Enterprise Coworker uses agentic CX orchestration grounded in brand, customer, and channel intelligence. The real frontier is not orchestration or models. It is treating brand judgment as a machine-readable, runtime-enforceable “taste contract,” including the constraint sets and refresh…
Policy Is Still Using 2019 Math. Two Price Tags Show Why.
Two numbers from recent weeks point to the same leadership failure. Policy keeps debating categories, while operators price operations in real units: desks onboard today, and days at sea tomorrow. The advantage goes to whoever refreshes the underlying assumptions fastest.
Hire the Phase, Not the Title: What the Mini-CEO CMO and the AI Role Boom Have in Common
Read the ChatGPT Work marketing page with the titles erased and it stops looking like a product page. It looks like an org chart. OpenAI describes a loop: customer insights and brand standards in, briefs and assets out, performance reports back in, recommendations for the next launch. The frontier…
Four Million Videos a Day: How Marketing Became a Specification Problem
Four million cinematic social videos a day is not a creative milestone. It is proof that marketing’s bottleneck is now specification: the “compiler” layer that translates intent into model-executable video plans, and the human acceptance seam that keeps brand meaning intact at scale.
Spectrum Becomes Statecraft: The Physical Layer Is the Next AI Asset
SpaceX exploring terrestrial spectrum triggered an $80B market swing. I keep coming back to the reaction: sophisticated capital priced connectivity permission and timing years ahead of deployment. For agentic AI, the constraint is rotating away from model quality and toward the physical and institu…
Define 'Good' Before You Ship the Agent
Teams ship ambitious enterprise agents, usage rises, dashboards look alive, and then renewal collapses. The missing piece is not governance or measurement. It is leadership sequencing: define “good” as an objective function and renewal-grade artifact before you wire the agent into the workflow.
The Evidence Artifact Your Institution Can Actually Stand Behind
I keep coming back to a mismatch I can feel in every institution that says it wants AI and then quietly refuses to let it decide anything. The fix is not another policy. It is a product: evidence contracts that make probabilistic systems legible to deterministic acceptance gates.
The Four Councils That Actually Move AI
Semafor’s Global Advisory Board split into Frontier, Capital, Builder, and Founders looks like a conference taxonomy. My read is that it actually matches how AI power is exercised right now, especially as Europe pivots from AI regulation theater toward compute and gigafactory scale.
Drones Cannot Replace the Planes. They Replace the Delay.
The thing the drone replaces is not the aircraft. It is delay. When you treat latency as the priced metric, you start seeing the same advantage pattern across defense pivots, distribution-based fintech, and AI-native workflows.
When Orchestration Ships as a Feature: The Distribution Stack Is Now the Moat
OpenAI’s GPT-5.6 Ultra treats multi-agent coordination as a shipped feature. That shifts the bottleneck from model quality to the distribution stack: orchestration, partner routing, and fresh data. If you cannot explain the route from buyer intent to completed transaction, you do not have a moat. Y…
The Storefront That Closes: HP and Salesforce Are Building the Operating Layer Moat
A customer chats. The assistant sounds fluent. Then nothing closes. HP and Salesforce are shipping the missing layer: reusable operating surfaces that connect conversation to real business state. That is the moat shift happening now.
Reuters Put Verification on a Product Page. Marketing Should Treat It Like One.
I was clicking through Reuters Agency and stopped on a simple shift I had not registered before: verification sits in the Solutions menu as a product. Not an ethos. Not a footer trust statement. A contractual capability. That placement changes the workflow question for marketing teams shipping clai…
The Distribution Default: Why the Interface That Finishes the Job Beats the Model That Answers Fastest
When model quality drifts toward table stakes, the competitive moat moves to the customer-facing interface that captures intent and turns it into proprietary context. Microagi’s free cleanings for recorded data are the robotics version of what marketing agents must build next.
When the Artifact Arrives Finished: ChatGPT Work and the New Acceptance Loop
ChatGPT Work feels like a marketing speed upgrade until you notice the real shift: the agent collapses drafting into finished multi-format artifacts, and the only remaining human step is deciding whether to ship. That acceptance gate becomes the new work primitive.
Why Agentic Commerce Needs Channels, Not Just Conversations
Salesforce is pushing Agentforce Commerce into general availability, with commerce agents designed to take shoppers from discovery through checkout and post-purchase. The surprising lesson: the chatbot is the surface. The durable advantage is the channel layer that carries commitment state across s…
Familiarity Made the Trust Gap. Now the Interface Has to Close It.
Gen Z did not reject phones. They complained about interfaces that never learn how users disagree. Reuters 2026 finds AI chatbot news trust at 44% among users, while overall news trust falls to 37%. My read: usage rises faster than trust because product teams assumed trust would follow capability.…
The 11% Problem Is a Product Spec, Not a Readiness Gap
I keep coming back to IBM’s “11% readiness” stat. Most coverage reads it as a human literacy gap. My read is different: the missing piece is the product instrument panel that makes agent autonomy observable, auditable, and re-approvable. Treat the control gap as a spec, not a training problem.
Domain Expertise Is the Agent Multiplier. Institutional Memory Is the Interface.
Across roughly 400,000 Claude Code sessions, Anthropic found domain expertise predicts how much useful work an agent produces per instruction. OpenAI’s Codex adoption shows the same bottleneck moving into departments with real domain fluency. When execution gets cheaper, power shifts to what your o…
When the Loop Runs Itself, Acceptance Becomes the Job
OpenAI’s autonomous chemistry demo and Adobe’s creative agent expansion look like pure capability leaps. But the real shift is leadership. When the loop plans, executes, and iterates, humans move from performing steps to designing acceptance: what the loop is allowed to do, what artifacts prove out…
AI Governance Is a Clock, Not a Briefing
I keep coming back to the failure mode boards miss: it is not that leaders lack AI knowledge. It is that governance artifacts decay faster than frontier systems and agent workflows evolve. The fix is a governance clock with decision triggers.
The 19-Day Gap: When Frontier Model Access Became a Policy Variable
Nineteen days can be the difference between “frontier advantage” and “ordinary output.” Anthropic suspended Claude Fable 5 on June 12 due to US export controls and nationality verification constraints, then restored it on July 1 with quota limits. The operational takeaway: model access is not a ste…
The Sameness Discount: Why Curation Is Becoming a Product Primitive
AI-assisted websites are starting to look alike, including underlying tells. If output is cheap and similarity rises, the market premium moves to managed variance. Semafor Intelligence is an early example of curation becoming infrastructure: anchored claims, preserved disagreement, and structured d…
Stop Asking If It's AI. Ask If It's Additive.
I keep coming back to a leadership failure mode in AI content governance: teams walk in asking, “Is this AI?” and end up building a better detector and a worse product. The cleaner decision rule is simpler: reward additive intent, demand semantic clarity, and let community judgment downrank what fa…
The Micro-Work Episode: Why Desktop Agents Need a Stopping Rule, Not a Chat Window
Gemini Spark on macOS turns desktop agents from “chat you keep alive” into workflow infrastructure. But the advantage only lands when you redesign work as micro-work episodes: bounded inputs, a spec, a named done artifact, and a stopping rule.
The Micro-Habit Contract: What Gemini Spark on Mac and Vercel eve Are Really Teaching Desktop Agents
A chat window is a place you visit. A desktop agent with file access is a place work happens. The frontier shift is not model quality. It is the engineering of daily, bounded work units that produce named artifacts in the tools where decisions live.
Shared Context Is the New Territory
A one-person agent workflow breaks the moment it leaves the builder’s laptop. The frontier right now is not smarter prompts. It is portable context: reusable, versionable “Skills” that turn team judgment into delegation-ready infrastructure. Control that layer, and you gain institutional leverage.
Agents Are Running the Revenue Loop. What Do You Own?
The revenue loop no longer waits for a Monday campaign calendar. Klaviyo moved Composer and Customer Agent into public beta, and Braze made Operator and Agent Console generally available. Together they show a new marketing product primitive: agentic, stateful decision policies that coordinate creat…
The Signature Hasn't Moved: Why Duty-Persistence Is the Real Agent Design Pattern
AI can draft faster than any human review cycle. Courts are now clear that the human duty does not move. That creates a new product design pattern for agentic workflows: preserve state, custody, and provenance from generation to signature.
When the Creative Loop Never Stops, Approval Stops Working
I keep coming back to a leadership shift almost nobody is naming directly: once the creative loop never stops, approval becomes the bottleneck. Jasper’s end-to-end GEO agent makes brand representation a managed, always-on surface. The real question is whether your team can govern that loop with ass…
Hollow Usage: When Agents Become the User, the Renewal Has to Defend Itself
Usage charts can look healthy when agents do the work. But renewal risk quietly rises when agents replace the human champion who translated the product into business value. The durable counter is evidence your product emits by default, as renewal-grade artifacts.
The Evidence Contract: What the EU's Border System Just Taught Every AI Builder
When the EU Entry/Exit System went live, queues exposed a classic mismatch: probabilistic execution runs into deterministic institutional gates. The fix is not better models. It is the evidence contract, a product-shaped set of logs, decision traces, and provenance that auditors and receiving syste…
When Compute Gets Financed, Marketing Becomes an Underwriting Job
I was pricing a pilot and realized the hard question was not what the model could do. It was whether we could sustain throughput and experience through real, year-two load. Apollo and KKR financing AI compute make that question the new marketing primitive: a capacity-backed promise you can underwri…
The Day-90 Test: Why Hardware Export Deals Live or Die on the Sustainment Loop
Units delivered are the easy metric. Influence comes from what survives day 90: the update cadence, training discipline, service loop, and iteration pathway that keep the buyer inside your operating system.
The Distribution Moat: How Meta's Agent and Shutterstock's Licensing Are Remaking Commercial Workflows
A customer asks for a product variant on WhatsApp late at night. Two years ago, the thread waited for humans and later rights clearance. Now Meta’s Business Agent and Shutterstock’s licensing-aware generative workflow sketch a new standard: channel-native action paired with provenance-safe creation…
Design for Wobble: Compute Reliability Is Becoming a Customer Experience Surface
A France heatwave warmed nuclear reactors enough that operators throttled output. RTE said the grid was secure. Both statements were true, and the contradiction felt like a preview: “always-on” AI agent promises lean on assumptions that capacity can stop guaranteeing. Marketing and customer-facing…
The Genre Was the Trust. AI Just Broke the Genre.
A testimonial landed in my inbox last week and I did not forward it. The cadence was too clean, the arc too neat. My read: customer trust is now a design constraint because AI can mimic high-trust genres on demand. Genre cues no longer protect you. The fruits test does.
The Objective Function Error: Why Tokenmaxxing Breaks Reality and How Problem Selection Fixes It
Some teams turned weekly token burn into a leaderboard or a performance proxy. Frontier coding agents make generation cheap, so the dashboard starts lying. My read is that tokenmaxxing breaks reality because it mistakes output proxy for decision reality. The fix is not “more instrumentation.” It is…
The Label Is Now Part of the Ad
I keep noticing that the label is no longer a back-office checkbox. As New York requires conspicuous synthetic-performer disclosure and Meta embeds AI info labels in the ad UI, provenance becomes a creative field with its own pipeline constraints.
The Adaptive Storefront: Why AI-Mediated Customers Demand a New Experience Infrastructure
AI assistants are now reorganizing the journey before the shopper ever hits your site. Adobe’s Analytics data suggests AI-referred visitors browse longer and convert more strongly. The strategic question is what happens after the click: the adaptive storefront middle layer that turns pre-sorted int…
Announcements Don't Move Ships. Underwriters Do.
Oil may slide or bounce on headlines, but ships wait for insurance sign-off. The same mechanism applies to AI supply chains: announcements move narratives, not workloads. Throughput moves only when third parties can underwrite the operational reality.
The Camp Industry Already Mapped Consumer Fragmentation. Now Agents Can Read It.
Fragmentation is not the enemy. The U.S. camp industry has treated it as a structured map of belonging for 150 years. Frontier agents can now read that map and pre-assemble identity-fit offers, shifting customers from search to assent.
Token Cost Is the Real Infrastructure. The Rest Is Real Estate.
Most boards think AI infrastructure means data centers and GPUs. But the real gate is the unit economics of the agentic execution loop: tokens consumed per outcome, retries, tool calls, and re-plans. Caching turns tokens into margin, and digital trade agreements can function like a cost discount fo…
When the Loop Runs Faster Than the Record
The debate about AI ad labels misses the frontier shift: creative loops can now iterate, test, and publish faster than humans can reconstruct an auditable archive. In 2026 politics, narrative coherence becomes an operator-owned discipline, not an external compliance fix.
The Agentic Command Layer: Why the Next OS Battle Isn't a Phone
The next operating system battle will not be a better phone. It will be about who routes your intent into action. Apple, Google, and OpenAI are converging on the same prize: an agentic command layer that executes across the apps you already use. The strategic pressure is now clear, and the leadersh…
The AI Century Is a Calendar Claim, and the Calendar Is Political
A humanoid robot finishes a half-marathon, US lawmakers draft a ban on Chinese robots in federal procurement, and conflict baselines hit a post-1946 high. The collision points to one mechanism: the calendar gap between capability and permission is now a capital allocation variable.
The Real Leadership Upgrade for AI Is Faster Disbelief
When frontier models land inside the same AWS procurement and production plumbing enterprises already run, the adoption story becomes a leadership story. The danger is not hallucination. It is the moment a confident AI-generated rationale feels right and a leader stops looking for reasons to be wro…
Waymo Just Turned a Fleet Liability Into Booked Power for AI Compute
Waymo just treated retired robotaxi batteries like booked capacity for the grid. Not a sustainability story. A procurement-grade reliability layer with delivery timelines. This changes how power gets marketed, underwritten, and planned for AI-adjacent compute.
The Stop Contract: Why the Pause Layer Is Becoming the Real Agentic Moat
Zigment, Waldo, and IrisAgent launched into adjacent customer-facing categories this week. The market story is agentic CX. The real product story is the stop contract: a legible, editable boundary that decides when the system should act and when it must route to a human. That decision boundary is b…
The Brand That Won't Close: Always-On Agent Presence Replaces the Destination
Gemini Spark reads like a quiet reframing of what a brand is: not a place you visit, but an agentic presence that operates around the clock under user direction. Pair that with Kaltura’s embeddable agentic avatars and the shift becomes operational. The marketing surface is changing from sessions to…
Agents Are Deployable. The Trust Budget Is What Compounds the Advantage.
Agents are no longer waiting on capability. In a single window, OpenAI, Asana, Microsoft, and Meta shipped deployable agent primitives with built-in legibility and audit hooks. The gap moved into the buyer: trust. Teams that design a trust budget end to end can compound autonomy without leakage. Te…
The Agent Productivity Gap Is a Measurement Problem
The best agent stack scores barely above 30 percent on real-world post-production tasks. The gap is not a model verdict. It is an instrumentation verdict. Scaffold your agents with trace fields, step-conditioned evaluation, and weekly outcome review so your team can learn instead of hope.
Compute Geography Is Becoming the Routing Layer for Agentic Workloads
Compute geography is quietly becoming the routing layer for agentic workloads. When water, power, and interconnect lead times dominate, “siting” stops being procurement trivia and becomes a reliability and token-economics loop.
The Verification Stack Is the New Engineering Platform
The question is no longer whether your agents can generate artifacts quickly. It is what operational system decides those artifacts are correct enough to ship. Cisco and AWS both point to the same shift: verification is becoming the platform layer for agentic software.
Marketing's Integration Debt Is Becoming a Product Category
Two launches, one pattern: marketing platforms are starting to sell the governed workflow loop that keeps approvals, frequency rules, provenance, and measurement invariant while AI frontends change. That is how integration debt becomes a product category.
The AI SDR Debate Is a Category Error. The Handoff Artifact Is the Product.
The AI SDR argument is stuck on autonomy versus human selling. The real bottleneck is later: what the agent hands to a human at the exact moment a buyer decides whether the risk is worth it.
Attention-Reallocation Velocity Is the AI KPI Finance Has Been Missing
Dell proves the infrastructure demand. A token blowup proves appetite. Neither proves AI spend converts into customer motion. Attention-reallocation velocity is the 60-day, cohort-based KPI bundle finance teams can actually defend.
The Handoff Is the Product: Why Continuity Infrastructure Is the New CX Moat
Chatbots can be fluent. The competitive question is what survives the handoff. Weave’s omnichannel receptionist and the Workday-Microsoft agent identity push reveal a new CX surface: continuity infrastructure. Measure state preservation, not just conversation speed.
The Citation Footprint: Why AI Discovery Is a Proof Architecture, Not a Keyword Game
ChatGPT Search, Perplexity, and Google AI Overviews are making discovery and verification one screen. The winning GTM asset is no longer keyword rank. It is a citation footprint: a structured claim-to-proof architecture that earns trust when risk becomes salient.
Robinhood Gave the Agent Its Own Account
Robinhood did not add an AI feature to a trading app. It introduced a new product primitive: an agentic account with scoped money, a preview gate, a live activity feed, and instant revocation. That shifts consumer finance from choosing products to deploying strategies run by agents, with execution…
Goal Mode Just Changed What 'Done' Means
OpenAI made Codex goal mode broadly available across the app, IDE, and CLI. Gartner simultaneously established enterprise AI coding agents as an evaluable category and named OpenAI a Leader. Read together, they flip the operating contract: “done” can no longer be a sentiment. It must be an inspecta…
Integration Debt: Why AI Output Fails at the Decision Layer
The model worked. The workflow did not. As agentic AI moves into live operations, failures increasingly happen at the seam between recognized intent and the next state the system is allowed to enter. Call it integration debt. It is auditable, fixable, and it shifts the frontier advantage from model…
The Background Agent Economy: Why Persistent Execution Is the New Operating Model
The most important agent in your company is the one you never open a chat window with. Gemini Spark runs continuously inside Gmail, turning the inbox into persistent execution. The advantage is no longer what the model can answer. It is what the loop can do, remember, and prove when no one is watch…
The Agent-Consumable Action Is the New Unit of Performance Marketing
Google’s AI Mode and Rakuten’s Mirai are collapsing performance marketing surfaces into agent-native decision points. The old unit of measurement, the link click, is eroding. The new requirement is to publish agent-consumable actions with acceptance gates and state-change instrumentation so agents…
Verification Latency: The Marketing Metric the FTC Just Created
Deployment is collapsing from months to weeks. In the same week, the FTC punished “active listening” claims that did not match the system behind the campaign. The new bottleneck is verification latency: the time to produce auditable proof that a live AI claim matches current capability. Build a pro…
Authenticity Is Now a Design Choice
Match Group’s synthetic persona layer lets teams test features before real users react. In parallel, an AI-generated story won a major literary prize amid detection controversy. Together they show the same shift: authenticity stops being a default property of the artifact and becomes a deliberate d…
The Autonomy Kernel: How Agent Identity Becomes the New Geopolitical Stack
A 30-page NIST concept paper on agent identity and authorization looks like enterprise plumbing. Read again, and it is a market-shaping standard: the unit is not agent capability, but legible identity, scoped authorization, audit proof, and non-repudiation. China is moving with a different philosop…
When Agent Deployment Becomes a ZIP File, the Game Moves Upstairs
Blueprint skills are becoming portable, runtimes are becoming ZIP-file endpoints, and quality loops are surfacing in platform layers. When agent infrastructure stops being a craft bottleneck, the scarce resource moves upward: evidence of correct outcomes in your environment.
The Conversion Test: What Cisco's Backlog and Layoffs Really Demand
Cisco’s $5.3B AI infrastructure backlog and simultaneous layoffs look contradictory to the public. The market read a single decision: reallocate capital and labor toward agentic workflows, then accept the short-term optics. The problem is measurement. This essay introduces the conversion test: dema…
Who Defines Safe Enough Wins the Stack
A Reuters report on US and Chinese AI guardrails points at a deeper shift: safety is becoming an evaluation interface. The leverage is not headline agreement. It is who designs the proof artifacts, who approves the reviewer, and whether those evidence chains are portable enough to become market inf…
The Agent Moat Is Proof
For two years the AI question was “who has the best model?” Agents change the ordering of priorities. Once agents edit code, touch infrastructure, spend budgets, and trigger workflows, the new operational question is whether you can prove what they did, why they did it, and whether they were allowe…
The AI Workforce Bifurcation: Redesign or Retreat
GM laid off over 10% of its IT department while hiring for AI-native development, data engineering, and agent workflow roles. The real signal is workforce bifurcation: AI advantage compounds only when leaders redesign decision rights, evaluation artifacts, and accountability close to the work.
The Real AI Jobs Crisis Is a Design Problem, Not a Trade Problem
Cisco’s $5.3B AI backlog alongside thousands of layoffs is not a proof that AI “replaces jobs.” It is a proof that organizations are accumulating integration debt: the gap between buying AI speed and redesigning work. The operator move is a job-family selector audit plus evidence gates before scali…
The Clearance-to-Delivery Gap
The Beijing summit showed CEOs traveling with the headlines. The chip agenda stayed elsewhere. The gap to study is clearance versus delivery, and the discipline is refusing assent until throughput is auditable.
Capital Is a Selector: India’s Outflows Are an AI-Readiness Signal, Not a Verdict
Foreign investors pulled about $21B from Indian shares in the first four months of 2026. The lazy take is “India is failing.” The operator take is sharper: capital is selecting for AI-readiness where deployment and proof are visible, measurable, and auditable.
Owning the Model Is Not Owning the System
A local Qwen MTP GGUF build hit a runtime assert that looks minor until you notice what it reveals: possession is not control. In AI, autonomy only exists when the system can recover acceptable output after a dependency breaks.
Bureaucracy, Not Chat, Is Becoming AI's First Trillion-Dollar Market
AI is redefining value not through chat or interface demos, but through embedded workflow disruptions in governance-heavy sectors. This piece reveals how procurement, finance close, licensing, and compliance are becoming AI's largest, most durable budget lines—and why the true payoff lies in cycle-…
Hajar Labs
When Product Security Is Your Brand: Anthropic's Claude Code Leak Rewrites Supplier Trust
When product security is your brand, a mundane packaging mistake becomes a board-level trust event. Based on public technical writeups and community reporting, the Claude Code incident appears to have involved an npm packaging error that exposed a source-map artifact associated …