Claude Code team member Thariq makes the case for replacing Markdown with HTML as the output format for AI agents: HTML packs in more information, is easier to share, and supports two-way interaction, while Markdown's editing advantage stopped mattering once he switched to making changes through prompts.
Why it matters: Claude Code team members explain why they use HTML instead of Markdown as the agent output format, and share prompts you can use as-is along with the scenarios they fit.
Boris Cherny, the creator of Anthropic's Claude Code, said in an interview at Sequoia AI Ascent that he went all of 2026 without writing a single line of code, merging dozens of PRs a day—150 in a single day at his peak—doing most of his work from his phone, with 5 to 10 sessions and hundreds of agents running at any given time, plus thousands more chewing through deep tasks overnight.
Why it matters: Boris Cherny walks through Claude Code's path from incubation to a billion dollars in revenue, and lays out his calls: programming is solved, SaaS moats are being flattened, and organizational process is where the real edge is.
After analyzing the architecture that surfaced in the Claude Code source leak, the author argues that its core isn't a secret algorithm but a while loop plus a tool dictionary in under 30 lines of Python, driven by stop_reason !
Why it matters: From the leaked source, the author distills 12 composable agent-engineering patterns and lays out a four-week path to get started, useful for checking your own implementation for gaps.
Permission Protocol · AI Agent Incident TrackerSelectedAI score8787
Baoyu walks through Claude Code's prompt caching mechanism to explain why quotas burn so fast, and lays out rules for saving tokens. He points out that caching only applies to prefixes, the main agent's cache window is 1 hour, and sub-agents' is 5 minutes. Reading from cache costs about one-tenth of recomputing, so frequent /clear actually triggers a full-price context rebuild. The rule of thumb: if the cache is still warm and the task hasn't changed, keep chatting; only start a new session when the cache has expired, the task has shifted, or there's too much context noise.
Why it matters: Starting from the prompt caching mechanism, this explains Claude Code's quota consumption and gives the criteria for deciding whether to continue a session or start over, plus configuration you can copy.
Based on the Claude Code source code that leaked unexpectedly last week, Drew Breunig mapped out how the system prompt is assembled: components fall into two categories—always included and conditionally included—and shift based on toggles like output_style, repl_mode, user_type_ant, skills_enabled, and mcp_connected.
Why it matters: The author breaks down the dynamic assembly logic behind Claude Code's system prompt, showing how conditional context engineering works in practice.
Thariq Shihipar, an engineer on Anthropic’s Claude Code team, summed up what the team learned from using hundreds of active Skills internally, sorting them into nine categories: library and API references, product validation, data acquisition and analysis, business process automation, code scaffolding, code quality and review, CI/CD and deployment, operations runbooks, and infrastructure operations.
Why it matters: Anthropic’s internal classification system for hundreds of Skills, along with its writing tips, can be adapted to help teams design their own Skills.
Meta 对齐总监 Summer Yue 让 OpenClaw 整理邮箱，先用小型模拟收件箱测试，随后切到真实收件箱，智能体开始删除所有超过一周的邮件。她从手机反复发送 Do not do that、Stop、STOP OPENCLAW 等停止指令，智能体仍继续删除，最终 200 多封邮件被永久删除。分析认为这是目标锁定失败，停止指令没有独立的打断通道，删除操作也没有确认门禁。
Anthropic's red-team research shows that Claude Opus 4.6 can find 500 high-severity vulnerabilities in mature open-source projects like GhostScript and OpenSC—some of which have been sitting there for decades.
Why it matters: Using an RCE case he reproduced himself, the author shows that once AI drives the cost of finding vulnerabilities down, unmaintained software becomes the real risk surface.
The author suggests using agent session logs to improve CLAUDE.md or AGENTS.md files: Claude Code stores sessions in ~/.claude/projects, while Codex stores them in ~/.codex/sessions—both in JSONL format but with different schemas.
Why it matters: The author works backward from agent session logs to figure out what to improve in CLAUDE.md, and has open-sourced a CLI that cuts search time from several minutes down to seconds.
1/1Thu
Thursday
Permission Protocol · AI Agent Incident TrackerSelectedAI score7676
Dragos’ investigation shows that attackers used Claude Code and OpenAI GPT-4.1 to target the OT environment of a Mexican water company. Claude Code handled broad discovery, identifying vNode industrial gateways, researching vendor credentials, generating password lists, and executing password spraying, while GPT-4.1 handled structured data analysis and Spanish-language output.
Why it matters: Dragos reconstructed the full chain of how attackers used Claude Code and GPT-4.1 to conduct reconnaissance and password spraying against a Mexican water utility’s OT environment, showing how AI was actually divided across the intrusion lifecycle.
Jesse Vincent has released an open-source tool called packnplay. With a single command — `packnplay run claude --dangerously-skip-permissions` — it spins up a pre-configured throwaway container to run a coding agent.
Why it matters: The author wraps all the tedious setup for running a coding agent in a container into one command, and also shares exactly how he handles credential conflicts with Claude Code.
Why it matters: The author ported Claude's SKILL.md system to the Codex CLI, with tool mappings and install instructions, so you can judge whether reusing Skills across models is feasible.
The author built the episodic-memory plugin for Claude Code so it can search past session logs. By default, Claude Code deletes the .jsonl session logs under ~/.claude/projects after one month; you can extend retention via cleanupPeriodDays in ~/.claude/settings.json.
Why it matters: The author turned Claude Code's session logs into semantically searchable episodic memory, so readers can judge for themselves how long-term context is preserved across sessions.
Anthropic rolled out its first-party Skills system simultaneously on Claude Code, Claude.ai, and the Claude API, and author Jesse Vincent quickly followed with a new version of Superpowers built on the official Skills.
Why it matters: Drawing on nearly a month of hands-on use, the author compares the official Skills with his own setup and lays out the trade-offs involved in migrating.
Author Jesse Vincent released Superpowers, a set of Skills built on Claude Code's new plugin system. Once installed, it injects a guiding prompt through the session-start hook, prompting Claude to proactively search for and use these Skills.
Why it matters: The author packaged his own coding-agent workflow into an installable Skill plugin, so readers can directly reuse his implementation flow from brainstorming to TDD.
The author walks through their full workflow with Claude Code: first isolating tasks with git worktree, then using a brainstorming prompt to make Claude ask only one question at a time and confirm the design in stages, and finally using a planning prompt to break the plan into small tasks and write them into docs/plans/.
Why it matters: The author splits Claude Code into two sessions—an architect and an implementer—and shares reusable prompts plus a git worktree approach for isolating tasks.