Skip to main content

AnthropicAI Coding Agent

Claude Code

Reads a repository, edits files, runs commands, and returns changes for review, from the terminal, IDEs, a desktop app, the web, and CI.

Read the GILD Review
Save a product your organization uses.Keep up with Claude Code without adding it to your stack.Visit Website
GILD score, leaders and teams+72 Strong positive55 accounts from leaders and teams take a side, out of 139 sources reviewed. Updated Oct 25, 2026. Use a task pilot to measure accepted work and review time.See the evidenceHow this figure is made
Operating stack
Agentic Software Development

The 30-Second Take

Claude Code pays off where a test suite can prove the agent's change, and the time it saves writing code moves to review.

Intercom's Brian Scanlan credits shared skills and context for tripling merged pull requests per R&D head in sixteen months. Spotify counts a 76% rise in pull requests as that much more to review.

Best fit
Bounded repository changes with automated checks.
Main constraint
Reviewer time, which has to grow with everything the agent hands back.
First move
Require the shell sandbox in managed settings and commit a shared CLAUDE.md before one squad gets Team seats.

Why

Every paid Claude plan includes Claude Code, and sessions start in auto mode with the shell sandbox off. Staff review before throughput rises.

Buying and deployment

What engineering leaders should verify before a broader rollout.

How you buy it

US$20

A month on Pro. Team seats are US$25, Max starts at US$100, Enterprise adds usage at API rates.

Source 1, Anthropic, Claude pricing: Claude pricing

Details

Included in every paid Claude plan. Pro US$20 per month, or US$17 billed annually as US$200 up front; Max from US$100 per month. Team US$25 per seat per month, or US$20 billed annually, with Premium seats at US$125 or US$100. Enterprise US$20 per seat per month billed annually, plus usage at API rates. Max 5x and Max 20x include US$100 and US$200 a month in Claude Platform API credits, and Team up to US$500 a month pooled. Those credits do not pay for interactive Claude Code in the terminal, IDE, desktop, or web; they cover claude -p and the Agent SDK only when run with an API key from the linked Console organization.

Cost drivers

About US$13

Per developer per active day, Anthropic's average for enterprise deployments.

Source 2, Anthropic, Costs & billing: Costs & billing

Details

Plan usage is shared with Claude chat and resets on five-hour and weekly windows. On API billing, Opus 5.5 costs US$4 input and US$20 output per million tokens, and standard cache reads cost US$0.20 per million; Anthropic reports enterprise deployments average about US$13 per developer per active day. Fast mode on Opus 5.5 runs up to 2.5x faster at US$8 input and US$40 output per million tokens, double the standard rate, with caching multipliers applied on top.

Where it runs

CLI, IDEs, Slack, CI

Plus a desktop app, the web, and mobile. On Team and Enterprise, Slack requests run as Claude Tag.

Source 3, Anthropic, Platforms & setup: Platforms & setup

Details

Terminal CLI, VS Code and JetBrains extensions, a desktop app, the web, the Claude mobile app, Slack, and GitHub Actions or GitLab CI. On Team and Enterprise, Slack requests run through Claude Tag, the organization's shared @Claude identity with admin-configured access.

What to verify

Training setting

Free, Pro, and Max data trains models when it is on. Team, Enterprise, and API data does not.

Source 4, Anthropic, Data usage: Data usage

Details

Free, Pro, and Max data trains models when the setting is on; Team, Enterprise, and API data does not. The seat allowance is the default ceiling; past it, usage credits run once an admin turns them on, with spend limits for the organization, a group, or a member. Code Review is billed separately from the plan.

Sourced from Anthropic documentation

Signals

What is the current evidence saying about Claude Code?

A source-linked read of what engineering leaders, teams, and builders report about Claude Code.

Whose sources count

GILD score, leaders and teams

+72 Strong positive net

of 55 sources that take a side. Accounts matched to a named person count more.

How this figure is made

Current sentiment

Positive, mixed and negative sources

Engineering leaders and teams report the clearest value after they establish shared conventions and context. Their main concerns are review load and the amount of work to delegate.

Mar 2025 to Oct 2026.

117 sources from practitioners, by month and sentiment.

What practitioners praise

Top praise of 5. Each mark is a source on it; filled marks praise it.

  1. Codebase UnderstandingAll 9 sources on it praise it

Most common concerns

Top concern of 5. Each mark is a source on it; filled marks raise it.

  1. Requires Strong Review15 of 22 sources on it raise it as a concern
All 139 sources behind these findings
  1. Code ReviewAnthropic (Official), Product documentation, checked Oct 25, 2026
  2. X post by @MaanavA51108Maanav Agrawal, founder building Memoir (self-described: Five parallel Claude Code chats on one product), X post, published Oct 21, 2026
  3. OverviewAnthropic (Official), Product documentation, checked Oct 12, 2026
  4. QuickstartAnthropic (Official), Product documentation, checked Oct 12, 2026
  5. Common workflowsAnthropic (Official), Product documentation, checked Oct 12, 2026
  6. How Claude remembers your projectAnthropic (Official), Product documentation, checked Oct 12, 2026
  7. Settings files and precedenceAnthropic (Official), Product documentation, checked Oct 12, 2026
  8. Configure permissionsAnthropic (Official), Product documentation, checked Oct 12, 2026
  9. AuthenticationAnthropic (Official), Product documentation, checked Oct 12, 2026
  10. SecurityAnthropic (Official), Product documentation, checked Oct 12, 2026
  11. Data usageAnthropic (Official), Product documentation, checked Oct 12, 2026
  12. Manage costs effectivelyAnthropic (Official), Product documentation, checked Oct 12, 2026
  13. Track team usage with analyticsAnthropic (Official), Product documentation, checked Oct 12, 2026
  14. MonitoringAnthropic (Official), Product documentation, checked Oct 12, 2026
  15. Hooks referenceAnthropic (Official), Product documentation, checked Oct 12, 2026
  16. Create custom subagentsAnthropic (Official), Product documentation, checked Oct 12, 2026
  17. Connect Claude Code to tools via MCPAnthropic (Official), Product documentation, checked Oct 12, 2026
  18. Use Claude Code in VS CodeAnthropic (Official), Product documentation, checked Oct 12, 2026
  19. JetBrains IDEsAnthropic (Official), Product documentation, checked Oct 12, 2026
  20. Claude Code GitHub ActionsAnthropic (Official), Product documentation, checked Oct 12, 2026
  21. Platforms and integrationsAnthropic (Official), Product documentation, checked Oct 12, 2026
  22. Configure the sandboxed Bash toolAnthropic (Official), Product documentation, checked Oct 12, 2026
  23. Model configurationAnthropic (Official), Product documentation, checked Oct 12, 2026
  24. Comment on: Meta and Microsoft take steps to reduce employee usage of Claude AIChris_Newton, Hacker News commenter (self-described: Company work on a long-lived web application, using API tokens), Hacker News comment, published Oct 8, 2026
  25. Comment on: Ask HN: Is anybody producing good code with coding agents?saluki, Hacker News commenter (self-described: Team using Claude Code in the terminal, with pull request review), Hacker News comment, published Oct 3, 2026
  26. Trying the Software Factory pattern.Will Larson, Chief Technology Officer at Imprint (self-described: Imprint, company-wide rollout across frontend, backend, infrastructure and data repositories), Personal blog post, lethain.com, published Sep 20, 2026
  27. Comment on: Launch HN: Skillsync (YC W26), AI chat sessions made portable across agentsbtown, Hacker News commenter (self-described: Team that shares Claude Code sessions between colleagues), Hacker News comment, published Sep 17, 2026
  28. Comment on: Debian votes to allow "responsible use of generative AI"jermaustin1, Hacker News commenter (self-described: Contract developer integrating code a client's managers generate with Claude), Hacker News comment, published Aug 29, 2026
  29. Comment on: Gemini Omni 1.1 Flashmh-, Hacker News commenter (self-described: Chief technology officer at a Google Cloud customer with an eight-figure annual commit), Hacker News comment, published Aug 27, 2026
  30. Breaking Claude Code Opus 5 Auto ModeJohann Rehberger, security researcher, Embrace The Red (self-described: Red-team test of Claude Code's default Auto Mode against a malicious website), Independent security evaluation, Embrace The Red, published Aug 26, 2026
  31. Comment on: Slop Debtkren87, former tech lead (self-described: Tech lead of a team measured by merge-request count and given fixed deadlines), Hacker News comment, published Aug 23, 2026
  32. AI Coding Benchmark: Claude Code vs CursorSedat Dogan and Şevval Alper, AIMultiple (self-described: Industry research firm; the same full-stack tasks run by CLI agents and AI code editors), Independent benchmark, AIMultiple, published Aug 21, 2026
  33. Comment on: AI Coding and Its Discontentsjaegerpicker, Hacker News commenter (self-described: Engineer with 25 years of experience; their team plans, codes, and verifies tickets with Claude), Hacker News comment, published Aug 12, 2026
  34. Comment on: Kimi K3-256kcybersec47, Hacker News commenter (self-described: Hard reverse-engineering and research tasks), Hacker News comment, published Jul 30, 2026
  35. I cancelled Claude after 6 months of daily use: here's what went wrongAnurag Singh, tech writer at XDA Developers (self-described: Own technical projects on the US$20 Pro plan), First-person review column, XDA Developers, published Jul 27, 2026
  36. Comment on: The new rules of context engineering for Claude 5 generation modelsthrowatdem12311, Only staff engineer at a company (self-described: Small company, half-million-line codebase, reviewing PRs from 3 offshore developers), Hacker News comment, published Jul 26, 2026
  37. Comment on: Who's afraid of Chinese models?aurareturn, Hacker News commenter (self-described: Company that standardized on Claude), Hacker News comment, published Jul 21, 2026
  38. Comment on: AI is a bad tooldeterministic, Maintainer of a large, business-critical C++ application (self-described: C++ application of over a million lines used by major international companies), Hacker News comment, published Jul 16, 2026
  39. Choosing the right tool safety approach for coding agentsRobat Williams, developer at Scott Logic (Scott Logic, software consultancy; hardening coding agent setups), Company engineering blog post, Scott Logic, published Jul 15, 2026
  40. Agentic Coding at ClickHouseAlexey Milovidov, Co-Founder and Chief Technology Officer at ClickHouse (ClickHouse engineering, large C++ codebase, company-wide agent use), Company blog post, ClickHouse Blog, published Jul 14, 2026
  41. Comment on: Write code like a human will maintain itradlad, Hacker News commenter (self-described: Workplace with about two engineers per service; runs several Claude Code sessions on unrelated work), Hacker News comment, published Jul 10, 2026
  42. Comment on: Godot will no longer accept AI-authored code contributionsmaiybe, Hacker News commenter (self-described: Team building a Godot mobile app, with shared Claude skills for writing and auditing pull requests), Hacker News comment, published Jul 1, 2026
  43. Comment on: Tokenmaxxing is dead, long live tokenmaxxinglinsomniac, Hacker News commenter (self-described: Company AI lead; most employees on US$20 monthly plans), Hacker News comment, published Jun 28, 2026
  44. Comment on: Claude Tagthewebguyd, Hacker News commenter (self-described: Company that rolled Claude Cowork out to all non-technical staff on Microsoft Teams), Hacker News comment, published Jun 23, 2026
  45. Comment on: Claude Tagisusmelj, Hacker News commenter (self-described: Team organization on an Anthropic plan), Hacker News comment, published Jun 23, 2026
  46. Scaling Myself: How I Run 22 Claude Code Sessions for DS4 Migrationcyan, iOS engineer on the Mercoin (MC) Client team (self-described: Mercari / Mercoin client team, design-system migration), Engineering blog post, Mercari Engineering blog, published Jun 22, 2026
  47. Comment on: Claude Code scans your whole drive, admits it when caughtteekert, Hacker News commenter (self-described: Organization that runs AI agents only in containers or scoped Linux users), Hacker News comment, published Jun 20, 2026
  48. Doubling the productivity of your engineering team using AI (Brian Scanlan)Brian Scanlan, Senior Principal Systems Engineer at Intercom (self-described: Intercom R&D organization), Conference talk (recorded, DX Annual), Engineering Enablement by DX (DX Annual talk; recap by Justin Reock), published Jun 15, 2026
  49. Agentic Testing: Where Agents Fit in the E2E Testing StackSergii Gorbachov, Staff Software Engineer at Slack (Slack end-to-end testing, more than 200 agent runs in test workspaces), Company engineering blog post, Slack Engineering, published Jun 11, 2026
  50. Comment on: Anthropic/OpenAI may be spending more than US$1000 for every US$100 you pay themjabart, Hacker News commenter (self-described: Team on a Claude Team plan, working in a codebase of millions of lines), Hacker News comment, published Jun 7, 2026
  51. A UI Framework Migration in the Age of AI: How We Migrated Svelte 4 to 5 with Claude Code in Six HoursPablo Arellano, Software Developer at Atolio (Atolio engineering team, Svelte front end with thousands of files), Company engineering blog post, Atolio Blog, published Jun 4, 2026
  52. Coding Is No Longer the Constraint: Scaling Developer Experience to Teams and Agents at SpotifyNiklas Gustavsson, Chief Architect and VP of Engineering at Spotify (Spotify engineering; Claude Code and the Honk background agent built on the Claude Agent SDK), Company engineering blog post, Spotify Engineering (Code with Claude 2026 talk), published Jun 3, 2026
  53. Comment on: Anthropic raises $65B in Series H funding at $965B post-money valuationtoraway, Hacker News commenter (self-described: Engineer who recommends the company's AI coding subscriptions), Hacker News comment, published May 29, 2026
  54. Codex vs Claude Code: Why I Switched for Daily DevelopmentKaran Goyal, Senior Shopify developer (self-described: Solo client work: Shopify apps, Next.js pages, SEO and content), Personal blog post, karangoyal.cc, published May 27, 2026
  55. How Salesforce Engineering Became Truly AgenticSrinivas Tallapragada, President and Chief Engineering and Customer Success Officer at Salesforce (Salesforce global engineering, Claude Code rolled out to every engineer), Company newsroom post, Salesforce News, published May 27, 2026
  56. Comment on: The Eternal Sloptemberjvidalv, Hacker News commenter (self-described: Ten-year-old codebase shared by more than 200 engineers, plus a five-engineer codebase), Hacker News comment, published May 25, 2026
  57. What I Learned Using Claude Code as an Engineering Manager During Focus WeekAhmad El-Shafey, Senior Engineering Manager at Slack (self-described: Slack, their engineering team's repo during Focus Week), Personal blog post, Medium, published May 13, 2026
  58. Why I Cancelled My Claude Code and ChatGPT Pro Subscriptions (for Side Projects)Akos Komuves, Developer with a day job, writing about side projects (self-described: Side projects, including HRS (a TUI time tracker)), Newsletter, Bitsy (Substack newsletter), published May 10, 2026
  59. Comment on: An AI agent deleted our production database. The agent's confession is belowsfink, Hacker News commenter (self-described: Work use with an employer-owned API key, sandboxed in firejail), Hacker News comment, published Apr 27, 2026
  60. Comment on: An AI agent deleted our production database. The agent's confession is belowdelegate, Hacker News commenter (self-described: Local database work outside production, with a backup taken first), Hacker News comment, published Apr 27, 2026
  61. Claude Code/Cowork monitoring at scale with Otel & ElasticSpencer Niemi, Elastic InfoSec (self-described: Elastic engineering organization; the InfoSec team monitors agent use with OpenTelemetry), Company engineering blog post, Elastic Security Labs, published Apr 25, 2026
  62. Comment on: An update on recent Claude Code quality reportsbmitc, Hacker News commenter (self-described: Company with tens of thousands of employees), Hacker News comment, published Apr 25, 2026
  63. Comment on: I cancelled Claude: Token issues, declining quality, and poor supportstan_kirdey, Hacker News commenter (self-described: Paid subscriber on the US$20 Pro plan, then the US$200 Max plan), Hacker News comment, published Apr 24, 2026
  64. Claude Code Tips I Wish I'd Had From Day OneCaroline Schneider, Marmelab (self-described: Marmelab, software agency; client projects and its react-admin and Atomic CRM open source frameworks), Company engineering blog post, Marmelab, published Apr 24, 2026
  65. Is Claude Code going to cost US$100/month? Probably not, it's all very confusingSimon Willison, Independent developer and blogger who teaches Claude Code (self-described: Pays for their own Max plan; runs coding agent tutorials), Personal blog post, simonwillison.net (Simon Willison's Weblog), published Apr 22, 2026
  66. Background Coding Agents: Supercharging Downstream Consumer Dataset Migrations (Honk, Part 4)Devon Edwards Joseph, Senior Engineer at Spotify (Spotify data team migrating downstream dataset consumers with Honk and Fleetshift), Engineering blog post, Spotify Engineering, published Apr 22, 2026
  67. Comment on: ChatGPT Pro now starts at US$100/monthsassymuffinz, Hacker News commenter (self-described: One-week trial on their own codebase; normally VS Code with occasional Copilot), Hacker News comment, published Apr 9, 2026
  68. Why I Cancelled Claude Max After 13 Months and What I'm Testing with Codex NextChandler Nguyen, self-taught builder (self-described: Solo, own products (STRATUM, DIALOGUE, course platform, personal site)), Personal blog post, chandlernguyen.com, published Apr 3, 2026
  69. [MODEL] Claude Code is unusable for complex engineering tasks with the Feb updatesStella Laurenzo (stellaraccident), software engineer (employer unnamed in the issue) (self-described: Compiler/GPU projects; the issue's first version says every senior engineer on the author's team reported similar experiences), GitHub issue, published Apr 2, 2026
  70. How we made Trail of Bits AI-native (so far)Dan Guido, CEO at Trail of Bits (Security consultancy; every employee on Claude Code with shared configs and sandboxes), Company blog post, Trail of Bits Blog, published Mar 31, 2026
  71. Comment on: Go hard on agents, not on your filesystemgurachek, Hacker News commenter (self-described: Solo, personal blog on Apache), Hacker News comment, published Mar 28, 2026
  72. What we're learning from engineers getting the most from Claude CodeAndrew Walker, Engineering Director at Fin.ai (Intercom) (self-described: Intercom engineering organization), Company blog post, Fin Ideas (Intercom), published Mar 25, 2026
  73. [BUG] Claude Max plan session limits exhausted abnormally fast since March 23, 2026 (CLI usage)karenrebecag, GitHub user (self-described: Max plan through the Claude Code CLI, same workload each day), GitHub issue on anthropics/claude-code, published Mar 24, 2026
  74. How we use Claude Code today at IntercomBrian Scanlan, engineer at Intercom (self-described: Intercom, company-wide Claude Code platform (engineering, data, support, product)), Company blog post, Fin Ideas (Intercom), published Mar 19, 2026
  75. Comment on: Ask HN: How is AI-assisted coding going for you professionally?alexmuro, Runs a small analytics and web products lab; manages 5 developers (self-described: Small lab building a React/Node analytics and GIS platform for clients), Hacker News comment, published Mar 15, 2026
  76. Feature Request: Persistent Memory Across Context Compactions (59 compactions, built our own)Haustorium12, GitHub user (self-described: 12 to 18 hours a day across two machines and six active projects), GitHub issue on anthropics/claude-code, published Mar 15, 2026
  77. Claude Code and COBOL modernization: What’s the reality?Shodhan Sheth, Enterprise Modernization, Platforms & Cloud Lead at Thoughtworks (Thoughtworks UK legacy modernization practice, 18 months of AI-assisted mainframe work, with Alessio Ferri), Company article, Thoughtworks Insights, published Mar 2, 2026
  78. Comment on: Ask HN: What's it like working in big tech recently with all the AI tools?cebert, Hacker News commenter (self-described: US$20 billion technology company with an enterprise subscription; TypeScript team), Hacker News comment, published Feb 26, 2026
  79. Comment on: Hugging Face Skillssothatsit, Hacker News commenter (self-described: Engineering team that wrote CLIs and skills for its internal systems), Hacker News comment, published Feb 24, 2026
  80. Comment on: How I use Claude Code: Separation of planning and executionraw_anon_1111, Hacker News commenter (self-described: US consulting company on an enterprise plan), Hacker News comment, published Feb 22, 2026
  81. How We Trained Our Engineering Team To Actually Use Claude Code (not just chat with it)Kat Korson, Director at Red Eagle Tech (self-described: Red Eagle Tech engineering team building bespoke software and an e-commerce SaaS platform), Company-authored engineering practice account, Collabnix, published Feb 21, 2026
  82. [BUG] [URGENT!!!] Claude Code is hanging / freezing / stuck on heaps of prompts for 5-20minutes or more.nullbio, GitHub user (self-described: Daily Claude Code use since the Opus 4.6 release), GitHub issue on anthropics/claude-code, published Feb 17, 2026
  83. Teaching Claude the Old Tricks: Lessons from Migrating Legacy Code to an AI-Assisted WorkflowCharles Herring, co-founder and chairman of WitFoo (self-described: WitFoo, cybersecurity platform; a decade of production code moving to a rebuilt product), Personal blog post, charlesherring.com, published Feb 10, 2026
  84. Claude Code: Production ready code in a two-week sprintChad Pytel, Developer and CEO at thoughtbot (thoughtbot consultancy team building a client Rails app), Company blog post, thoughtbot blog, published Feb 9, 2026
  85. What our web team learned using Claude Code for a monthJuwan Wheatley, Expo engineer (self-described: Expo web team and company-wide Claude Code users), Company engineering blog post, Expo Blog, published Feb 6, 2026
  86. My AI Adoption JourneyMitchell Hashimoto, Ghostty (self-described: Ghostty terminal and his own projects), Personal blog post, mitchellh.com, published Feb 5, 2026
  87. Comment on: Auto-compact not triggering on Claude.ai despite being marked as fixedbtown, Hacker News commenter (self-described: Uses Claude Code and encourages their team to use it), Hacker News comment, published Jan 23, 2026
  88. Cloud Coding Agents at HubSpotStephan Lensky, senior engineer on the AI Developer Experience team at HubSpot (HubSpot engineering, self-hosted cloud agent platform (Crucible) on Kubernetes), Company engineering blog post, HubSpot Product Blog, published Jan 23, 2026
  89. Don't fall into the anti-AI hypeantirez, Redis developer (self-described: Redis and his own C libraries), Personal blog post, antirez.com, published Jan 11, 2026
  90. [Bug] Excessive token usage in Claude Code 2.1.1 - 4x+ faster rate consumption than previous versionsh3d0, GitHub user (self-described: Max plan; the same projects and routine for months), GitHub issue on anthropics/claude-code, published Jan 8, 2026
  91. Claude Code and What Comes NextEthan Mollick, One Useful Thing (self-described: Nontechnical user building and launching a small web business), Newsletter post, One Useful Thing (Substack), published Jan 7, 2026
  92. Comment on: Scaling LLMs to Larger Codebasesgodzillafarts, Hacker News commenter (self-described: Engineering team that adapted HumanLayer's context-engineering prompts for its own stack), Hacker News comment, published Dec 22, 2025
  93. Comment on: Ask HN: How can I get better at using AI for programming?thyb23, Hacker News commenter (self-described: Max plan with Opus 4.5, after trying Gemini CLI, Cursor, Codex CLI, and Antigravity), Hacker News comment, published Dec 14, 2025
  94. Comment on: The 7 Habits of Highly Ineffective Agentstobyhede, Hacker News commenter (self-described: Solo side project, Rust game with Bevy), Hacker News comment, published Dec 13, 2025
  95. Background Coding Agents: Predictable Results Through Strong Feedback Loops (Honk, Part 3)Max Charas and Marc Bruggmann, Senior Staff and Principal Engineers at Spotify (Spotify Fleet Management, background agent (Honk) running Claude Code across thousands of components), Engineering blog post, Spotify Engineering, published Dec 9, 2025
  96. Comment on: Has the cost of building software dropped 90%?loandbehold, Remote developer (self-described: Remote developer on a team that uses GitHub Copilot), Hacker News comment, published Dec 8, 2025
  97. No Vibes Allowed: Solving Hard Problems in Complex Codebases - Dex Horthy, HumanLayerDex Horthy, Founder of HumanLayer (self-described: HumanLayer, three-person engineering team), Conference talk (AI Engineer Code 2025), YouTube (AI Engineer channel; transcript on ai.engineer), published Dec 2, 2025
  98. What I learned building an opinionated and minimal coding agentMario Zechner, Independent developer, coach and speaker (self-described: Solo; built their own open source coding agent (pi) after using Claude Code), Personal blog post, mariozechner.at, published Nov 30, 2025
  99. Comment on: Anti-patterns while working with LLMshamasho, Hacker News commenter (self-described: Team sharing several AWS development instances), Hacker News comment, published Nov 28, 2025
  100. Background Coding Agents: Context Engineering (Honk, Part 2)Max Charas and Marc Bruggmann, Senior Staff and Principal Engineers at Spotify (self-described: Spotify, fleet-wide background coding agent for migrations), Engineering blog post, Spotify Engineering, published Nov 24, 2025
  101. Comment on: Microsoft AI CEO pushes back against critics after recent Windows AI backlashewoodrich, Hacker News commenter (self-described: Team whose manager asked for one AI coding subscription for everyone), Hacker News comment, published Nov 20, 2025
  102. X post by @IceSolstsolst/ICE (@IceSolst), security engineer (self-described: Overnight Claude Code loop on a detailed task list, paid with employer tokens), X post, published Nov 15, 2025
  103. Rolling out Claude Code at AirtaskerAsher Moore, engineering leader at Airtasker (self-described: Airtasker, engineering-wide rollout via AWS Bedrock), Engineering blog post, Technology @ Airtasker (Medium), published Nov 10, 2025
  104. How I Use Every Claude Code FeatureShrivu Shankar, leads AI developer tooling at Abnormal AI (self-described: Abnormal AI engineering, shared monorepo), Personal blog post, Shrivu's Substack (blog.sshh.io), published Nov 2, 2025
  105. Claude Code Can Debug Low-level CryptographyFilippo Valsorda, open source maintainer (self-described: Go standard library cryptography; a new ML-DSA implementation), Personal blog post, words.filippo.io, published Nov 1, 2025
  106. Living dangerously with ClaudeSimon Willison, Independent developer and blogger (self-described: Solo, side projects and research experiments), Annotated talk on personal blog, simonwillison.net (Simon Willison's Weblog), published Oct 22, 2025
  107. [BUG] ## CRITICAL: Claude Code executed rm -rf deleting entire home directoryMike Wolak, developer (self-described: Own firmware project on Ubuntu under WSL2), GitHub issue, published Oct 21, 2025
  108. Comment on: Claude Code on the webkelvinjps10, Hacker News commenter (self-described: Terminal work after trying Codex; vim prompt editing and shell tools), Hacker News comment, published Oct 20, 2025
  109. Just Talk To It - the no-bs Way of Agentic EngineeringPeter Steinberger, Independent developer (self-described: Solo, ~300k LOC TypeScript React app plus extension and CLI), Personal blog post, steipete.me, published Oct 14, 2025
  110. [Meta] Unexpected change in Claude usage limits as of 2025-09-29. (30+ reports)emcd, GitHub user (self-described: Claude Pro subscriber using Claude Code daily), GitHub issue on anthropics/claude-code, published Oct 7, 2025
  111. How I'm using coding agents in September, 2025Jesse Vincent, Independent developer (self-described: Solo, 3-4 parallel projects on a single codebase), Personal blog post, blog.fsck.com (Massively Parallel Procrastination), published Oct 5, 2025
  112. 90%Armin Ronacher, founder of a new company (self-described: New company; a 40,000-line Go email service with generated Python and TypeScript SDKs), Personal blog post, lucumr.pocoo.org, published Sep 29, 2025
  113. Claude Code has changed how we do engineeringJonathan Pedoeem, Technical founder of PromptLayer (self-described: PromptLayer, startup engineering team), Company blog post, PromptLayer blog, published Sep 3, 2025
  114. First attempt will be 95% garbage: A staff engineer's 6-week journey with Claude CodeVincent Quigley, then Staff Engineer at Sanity (self-described: Sanity, engineering team (internal workshop write-up)), Engineering blog post, Sanity Engineering, published Sep 2, 2025
  115. Comment on: Survey: a third of senior developers say over half their code is AI-generatedstavros, Hacker News commenter (self-described: Own development work on the Max plan; has it write small functions that stavros then composes), Hacker News comment, published Sep 1, 2025
  116. Practical Lessons Learned using Claude Code to automate IntegrationsPeter Silberman, Chief Technology Officer at Fixify (Fixify, IT operations company with its own integration SDK), Newsletter post (Substack), published Aug 20, 2025
  117. Testing AI coding agents (2025): Cursor vs. Claude, OpenAI, and GeminiMitch Alderson, Render (self-described: Render engineering; timed tests on its own production codebases and a new Next.js app), Independent benchmark, Render engineering blog, published Aug 12, 2025
  118. Getting Good Results from Claude CodeChris Dzombak, software engineer (self-described: Personal projects, about 12 programs), Personal blog post, dzombak.com, published Aug 8, 2025
  119. My One-Month Journey with Claude CodeEric Cheng, Software engineer (self-described: Large refactoring project involving thousands of files), Personal blog post, DEV Community, published Aug 6, 2025
  120. Adopting Claude Code: Riding the Software Economics SingularityMaxime Beauchemin, Founder and Chief Executive Officer of Preset (Preset, the company behind Apache Superset; large open source repository), Company blog post, Preset Blog, published Aug 4, 2025
  121. Comment on: 6 weeks of Claude Codefdsf111, Hacker News commenter (self-described: Developer at a company, Python/JavaScript work), Hacker News comment, published Aug 2, 2025
  122. Claude Code Memory Leak - Process Grows to 120+ GB RAM and Gets OOM Killedchatgpt-copypasta, GitHub user (self-described: Ubuntu 24.04 machine with 128 GB of RAM, long coding sessions), GitHub issue on anthropics/claude-code, published Aug 1, 2025
  123. 6 Weeks of Claude CodeOrta Therox, engineer at Puzzmo (self-described: Puzzmo, small engineering team, monorepos), Company engineering blog post, Puzzmo Blog, published Jul 30, 2025
  124. How I got Claude to write code I could actually shipKush Gupta, engineer at Graphite (self-described: Graphite engineering team, monorepo), Engineering blog post, Graphite blog, published Jul 24, 2025
  125. Comment on: My experience with Claude Code after two weeks of adventuressankalp (dejavucoder), Hacker News commenter (self-described: Solo, large open source Ruby + TypeScript codebase), Hacker News comment, published Jul 17, 2025
  126. My Experience With Claude Code After 2 Weeks of Adventuressankalp (dejavucoder), Independent developer and AI engineering consultant (self-described: Gumroad bounties and consulting; medium Python and large Ruby/TypeScript codebases), Personal blog post, sankalp.bearblog.dev (sankalp's blog), published Jul 17, 2025
  127. How I Use Claude Code to Ship Like a Team of FiveKieran Klaassen, General manager of Cora at Every (self-described: Building Cora, an AI email assistant, as its hands-on developer), Newsletter essay (partly paywalled), Every (Source Code), published Jul 16, 2025
  128. How I use Claude Code (+ my best tips)Steve Sewell, Builder.io (self-described: Builder.io, large React codebase; a Cursor power user for a year before the switch), Company blog post, Builder.io, published Jul 11, 2025
  129. I Shipped a macOS App Built Entirely by Claude CodeIndragie Karunaratne, Engineer with a day job, building a side project (self-described: Solo side project, native macOS app (Context)), Personal blog post, indragie.com, published Jul 1, 2025
  130. How we're shipping faster with Claude Code and Git WorktreesRory Bain, Product Engineer at incident.io (self-described: incident.io, product engineering team), Engineering blog post, incident.io blog, published Jun 27, 2025
  131. Agentic Coding RecommendationsArmin Ronacher, Independent developer (self-described: Solo, own apps and open source library work), Personal blog post, lucumr.pocoo.org (Armin Ronacher's Thoughts and Writings), published Jun 12, 2025
  132. How Two Engineers Ship Like a Team of 15 With AI AgentsKieran Klaassen and Nityesh Agarwal, Cora team at Every (self-described: Every, two-person Cora engineering team), Podcast episode (recap page with video), Every (AI & I podcast), published Jun 11, 2025
  133. Field Notes From Shipping Real Code With ClaudeDiwank Singh, Julep (self-described: Julep, AI workflow orchestration; a backend with years of accumulated decisions), Personal blog post, diwank.space, published Jun 7, 2025
  134. Comment on: Cursor 1.0lmeyerov, Hacker News commenter (self-described: Solo; wrote an internal guide to their vibe-coding process), Hacker News comment, published Jun 5, 2025
  135. Claude Code is My ComputerPeter Steinberger, Independent developer (self-described: Solo, own Mac and personal projects), Personal blog post, steipete.me, published Jun 3, 2025
  136. Comment on: Ask HN: Cursor or Windsurf?erenst, Hacker News commenter (self-described: Solo, game development experiments; backend web development at work), Hacker News comment, published May 12, 2025
  137. Basic Claude CodeHarper Reed, CEO of 2389 Research (self-described: Own projects; notes their team follows a similar process), Personal blog post, harper.blog (Harper Reed's Blog), published May 8, 2025
  138. Nobody Codes Here AnymoreAlex Ghiculescu, co-founder of Tanda (self-described: Tanda, about 40 developers, 12-year-old Ruby on Rails SaaS codebase), Newsletter, Alex Ghiculescu's Newsletter (Substack), published Apr 29, 2025
  139. Claude Code saved us 97% of the work on the first try. Then it failed utterly.Alessio Ferri, Thoughtworks (self-described: Thoughtworks team building the CodeConcise legacy-modernization tool), Engineering blog post, Thoughtworks Insights blog, published Mar 10, 2025
The full readExplore all Claude Code Signals7 praise themes, 8 concerns, 5 points of agreement and 3 disagreements across all three groups, with all 139 sources open to inspect.

Where It Fits

Where does Claude Code belong in the stack?

Claude Code is Core across Build, supports Plan and Validate, and touches Ship and Operate.

Business area
Product & Engineering
Function
Software Development
Operating stack
Agentic Software Development
Primary role
Coding & Implementation Agent
Claude Code beside the rest of its stackThe Stack Field, every product on every layer
FootprintClaude Code across the Agentic Software Development stack. Bar height is its role on each layer.

  1. Plan

    1. Work planning: Supports

      Work planning, supports. Plan mode reads the code and proposes steps before any edit. The task itself stays in Linear, Jira, or GitHub Issues, which Claude Code reads through a CLI or an MCP server.

  2. Build

    1. Codebase context: Core

      Codebase context, core. It searches and reads the repository on demand, and loads CLAUDE.md files at the start of every session for build commands and conventions.

    2. Coding: Core

      Coding, core. Claude Code makes multi-file changes in an existing codebase from the terminal, an IDE, the desktop app, or a cloud session.

    3. Agent runtime: Core

      Agent runtime, core. Engineers steer it from the CLI, the VS Code and JetBrains extensions, the desktop app, the web, Slack, and GitHub, with permission modes from read-only Manual to auto.

    4. Tool execution: Core

      Tool execution, core. It runs shell commands under permission rules and calls tools through MCP servers. Admins can restrict both with managed settings.

  3. Validate

    1. Debugging: Supports

      Debugging, supports. It reproduces failures, reads logs and stack traces, and runs commands to test a fix. Error data from Sentry or another tracker arrives only through an MCP server or pasted context.

    2. Testing: Supports

      Testing, supports. It writes tests and runs your suite locally; a Stop hook can block a turn from ending until the suite passes. CI remains the gate.

    3. Code review: Supports

      Code review, supports. Code Review, for Team and Enterprise, posts severity-ranked inline comments on GitHub pull requests and bills tokens through usage credits. A review never approves or blocks a merge, so human approval stays the gate.

    4. Application security: Touches

      Application security, touches. It can review a diff for security issues on request. It does not scan dependencies or run SAST, so keep your application-security tools.

  4. Ship

    1. Source control: Touches

      Source control, touches. It creates branches, commits, and pull requests, and runs each parallel session in its own Git worktree. GitHub or GitLab stays the record.

    2. CI/CD: Touches

      CI/CD, touches. It runs inside GitHub Actions and GitLab CI when tagged with @claude, and can edit pipeline files. It does not deploy.

  5. Operate

    1. Observability: Touches

      Observability, touches. It reads production context only through MCP servers or pasted logs, and collects no telemetry from your services.

Market mapThe layers Claude Code covers, and who else owns each one.

  1. Codebase context & intelligence

    CoreBuild

  2. Coding & implementation

    CoreBuild

  3. Agent interface & developer runtime

    CoreBuild

  4. Tool execution & integrations

    CoreBuild

  5. Work planning & task decomposition

    SupportsPlan

7 more layers on the full map. Core, Supports and Touches describe coverage. They say nothing about quality.

Open Where It Fits: the Stack Field and the full map

Operator Insights

What are engineering leaders learning from deploying Claude Code?

See all 3 operator accounts

Comparisons

How does Claude Code compare with the alternatives?

  1. Claude Code vs Codex

    • AI Coding Agents

    Choose Codex to govern a mixed fleet from one console, with native Windows sandboxing and per-group policy. Choose Claude Code for a single-policy Mac and Linux fleet, at a higher token cost.

  2. Claude Code vs Cursor

    • AI Coding Agents

    Choose Cursor to pick a model per task in one editor, where the picks set the bill. Claude Code keeps every editor, JetBrains included, on Anthropic's models and a per-seat allowance; its VS Code extension also runs inside Cursor.

  3. Claude Code vs GitHub Copilot

    • AI Coding Agents

    Choose Copilot when delegated work should start in GitHub, where its cloud agent turns an issue into a pull request. Choose Claude Code when it starts at an engineer's terminal or IDE; it reaches GitHub through its GitHub app and Actions.

See all 13 comparisons

More on Claude Code

  1. GILD ReviewGILD's evaluation, with the buying facts
  2. Research3 pieces
  3. Guides4 guides
  4. Operator Insights3 operator accounts
  5. Where It FitsOn the Agentic Software Development stack
  6. Related Products8 products