Field guide / sub-domain
46
sources in this field
Updated July 26, 2026
Current thesis
The shortest path to orientation.
Claude Code generates $2.5B annually with 4% of GitHub commits. Anthropic's upstream Platform strategy includes Claude Managed Agents (PaaS, $2.58 per $1k service), the Advisor pattern (Opus planning + Sonnet/Haiku execution), and Monitor tool for event-driven scripts. First-party documentation spans a 14-minute video and 33-page Skills guide. Skill distribution has formalized around gstack (dotfiles for AI), ByteRover (queryable knowledge index), and self-improving loops graded against binary criteria. A GTM-specific stack chains 8 MCPs—Apollo, InstantlyAI, HeyReach, Findymail, Firecrawl, Gmail, Slack, HubSpot—enabling direct execution from sourcing through CRM management without human intermediaries. Enterprise adoption is accelerating: Intercom built 13 plugins with 100+ skills, and $1M+ customers grew 40x to 500+. NVIDIA hosts ~80 free models via APIs. Tolaria's MCP server lets Claude natively edit 100K+ LOC vaults. Claude Ads cuts waste 35% via 250+ audits. The ecosystem demonstrates a shift from chat-based research toward autonomous execution.
Evidence board
Claims worth carrying forward01A GTM stack for Claude Code combines 8 MCPs: Apollo (TAM sourcing/stakeholder mapping), InstantlyAI (email campaigns), HeyReach (multi-account DM prospecting), Findymail (email sourcing/verification), Firecrawl (web search/scrape), Gmail (thread search/drafting), Slack (context surfacing), and HubSpot (CRM object management and pipeline queries).
02Prediction: within a year, GTM teams will do the majority of their work through AI agents/Claude, shifting from manual tool-switching to agent-orchestrated workflows across sourcing, outreach, and CRM.
03Anthropic removed over 80% of Claude Code's system prompt for Claude Opus 5 and Fable 5 with no measurable loss on coding evaluations, reflecting a broader shift from rigid rules to model judgment.
04Old rule 'default to writing no comments, never write multi-line docstrings' has been replaced with 'write code that reads like the surrounding code: match its comment density, naming, and idiom' — newer models handle nuanced judgment calls without brittle guardrails.
05Giving Claude tool-usage examples can over-constrain newer models to a narrow exploration space; better to invest in tool/interface design (e.g., enum status fields like pending/in_progress/completed) that implicitly signals correct usage.
06Progressive disclosure now extends beyond skills to tools: some tools use 'deferred loading' requiring Claude to call ToolSearch to find full definitions, letting harnesses support far more tools without bloating context.
07CLAUDE.md and Skill.md files should not be exhaustive repositories of every practice; instead use a tree of files loaded progressively at the right time, keeping the root file lightweight and focused on non-obvious codebase 'gotchas.'
08Claude Code has moved from CLAUDE.md-based manual memory (# hotkey) to auto-memory, where Claude automatically saves relevant memories without explicit user instruction.
Adjacent fields
Latest evidence
Fivos Aresti
A GTM-focused Claude Code stack composed of 8 Model Context Protocol integrations covering sourcing, outreach, enrichment, scraping, email/Slack context, and CRM management, with a prediction that AI agents will handle most GTM work within a year.
Thariq
Anthropic's Thariq details how prompting and context engineering practices have shifted for Claude Opus 5/Fable 5: over 80% of Claude Code's system prompt was removed with no eval loss, as newer models handle judgment, tool design, and progressive disclosure better than rigid rules, repeated examples, or upfront specs.
Richard Seroter
Anthropic's Deputy CISO Jason Clinton details the security architecture behind an AI-native SDLC where Claude authors ~80% of merged code, describing four strategies: shifting security left, hard identity/access boundaries, layered automated+agentic review, and human-in-the-loop at high-leverage points.
Ruari Wood-Freeman
A GitHub-based plugin marketplace pattern allows centralized management of multiple Claude plugins with sub-skills and agents across an entire organization. Updates to the central repo automatically sync to all users' Claude instances, eliminating manual distribution overhead.
elvis
A workflow pattern combining LLM Wikis (structured knowledge stores) with interactive HTML Artifacts creates a bidirectional agent-UI layer. Artifacts are built on top of wikis, can invoke agents, and agents can update artifacts—enabling inbox-zero automation, research scheduling, topic discovery, and live figure generation from a single HTML file.
Suryansh Tiwari
Claude Code's real power comes from building the ecosystem around it—persistent memory (CLAUDE.md), project initialization (/init), parallel Git worktrees, MCP servers, plugins, subagents, and CI/CD integration—rather than prompting better. The claude-code-setup plugin automates recommending and installing these components.
Claude
Claude Code has launched 'agent view,' a consolidated list of all active sessions, currently available as a research preview. The feature addresses session management as Claude Code's multi-agent usage scales.
Yi Lu
claude-smart is an open-source plugin for Claude Code, Codex, and OpenCode that turns corrections and successful patterns into structured, reusable skills — stored locally at ~/.reflexio/ — rather than passive memory transcripts. It claims ~3× better rule retention than claude-mem, 70%+ reduction in planning iterations on repeated tasks, and runs entirely on-device with ONNX embeddings (all-MiniLM-L6-v2) plus hybrid vector+BM25 search.
Dan McAteer
'Understand Anything' is an open-source plugin for Claude Code, Codex, and OpenCode that analyzes a directory and builds an interactive knowledge base explaining structure rather than showing it. Notably, it generalizes beyond code — running /understand on an 888-episode Obsidian podcast vault (144K lines of markdown) produced a searchable knowledge graph.
Sean Geng
Sean Geng packaged @goodalexander's iterative plan-improvement technique as an installable Claude Code skill. The skill treats planning as a search problem: generate a plan, score it 0-100 against a rubric, critique specific weaknesses, rewrite, keep the best, and stop when scores plateau over a sliding window. One curl command installs it into ~/.claude/skills/plan-optimizer/.