Steven Gonsalvez

Software Engineer


CEREBRO


machine-read, human-curated

Coding agents

  • Claude Code v2.1.207 drops Auto mode's opt-in gate on Bedrock/Vertex/Foundry and fixes terminal freeze on long streamed output, useful if you're on managed cloud deploys.
  • Claude Code desktop ships in-app browser, sandboxed, lets agent read/click/interact with docs or dev servers directly, no more copy-paste context.
  • theo: Codex's "ultra" effort setting cascades to all subagents (token burn), no per-agent override; Claude Code lets you pin subagent effort independently.
  • theo again: Claude Code's subagent orchestration beats Codex's, programmatic orchestration still the only sane pattern for multi-agent coding.
  • theo and skirano on model routing: gpt-5.6-sol performs better inside Claude Code than Codex; skirano's stack = Sol Ultra for planning, Sol Medium for coding, Terra High for context/search subagents, Luna for trivial file ops.
  • rasbt: for agentic coding, prefer a cheaper model at higher effort over a pricier model at lower effort, same/better perf for less cost.
  • Someone built a phone-resumable Claude Code session via browser control panel, close laptop, pick up thread on mobile.

Agentic SaaS

  • Anthropic launched Claude Tag for Slack: tag Claude in a channel, it gets its own sandbox, clones repos, writes/tests/compiles code, sandbox destroyed after (details, proactive behavior here); beta live today for Enterprise/Team (more). 65% of Anthropic's internal product-team code now comes from this internal tool (bcherny).
  • MagicPath 2.0: multiplayer canvas where humans + coding agents (Codex, Claude Code) design/build together in real time against your codebase.

Coding agent tooling / MCP / skills

  • codebase-memory-mcp: tree-sitter-based code intelligence MCP server, indexes repos (Linux kernel in 3min), sub-ms structural queries, 99% fewer tokens, single static binary, no deps, big win for agent context efficiency.
  • mattpocock/skills: lightweight, composable Claude skills for real engineering (not vibe coding), model-agnostic, contrasts with heavier frameworks like GSD/BMAD/Spec-Kit.
  • addyosmani/agent-skills: production-grade skill set covering DEFINE→PLAN→BUILD→VERIFY→REVIEW→SHIP pipeline for coding agents.
  • google-labs-code/stitch-skills: Agent Skills standard library for Google Stitch, works across Antigravity, Gemini CLI, Claude Code, Cursor.
  • mattpocockuk floats shipping his own skills-installer CLI (npx @ai-hero/skills init) for per-agent customization and dependency tracking between skills.
  • SnapState: persistent state layer for AI agent workflows.
  • system_prompts_leaks: regularly-updated dump of leaked system prompts across Claude, GPT-5.6, Gemini, Grok, Cursor, Copilot, useful for prompt-engineering reverse-engineering.

CLI/TUI

  • claude-code-templates: ready-made agents/commands/hooks/MCP configs for Claude Code, plus new web dashboard (aitmpl.com) for browsing/installing.

Vibe-coding & infra

  • OpenMontage: open-source agentic video production system, 12 pipelines/52 tools/500+ skills, turns coding agent into a video studio, multi-agent (Claude/ChatGPT/DeepSeek) collab via Bloome.
  • NVIDIA congratulates xAI's Grok 4.5, trained on GB300 NVL72 systems, positioned for coding/agentic workloads, another entrant in the agentic-coding model race.

Signals

  1. v2.1.207
  2. davila7/claude-code-templates: CLI tool for configuring and monitoring Claude Code
  3. Tag Claude in a channel, it spins up an instance with its own sandbox. It clones repos, writes code, tests, compiles all in that isolated e…
  4. That's it. Point it at a channel, give it a task, and let it work. It's in beta on Slack today for Claude Enterprise and Team customers. Mo…
  5. Claude is really proactive with Claude Tag. You don’t need to prompt it to do work, it can do work proactively based on your instructions.…
  6. Introducing MagicPath 2.0. MagicPath is now a multiplayer canvas for humans and agents like Codex or Claude Code to design and build with A…
  7. If you set gpt-5.6-sol to "ultra", all the subagents it spawns will also be set to ultra. IMO this is a fumble. Causes massive token burn f…
  8. RT @ClaudeDevs: Claude Code on desktop now has an in-app browser. Claude can pull up docs, designs, or any other site. It can read, click t…
  9. I built a browser control panel for Claude Code — close your laptop, resume the session on your phone
  10. gpt-5.6-sol is meaningfully better in Claude Code than in Codex I'm going to crash out so badly over this
  11. DeusData/codebase-memory-mcp: High-performance code intelligence MCP server. Indexes codebases into a persistent knowledge graph — average…
  12. mattpocock/skills: Skills for Real Engineers. Straight from my .claude directory.
  13. SnapState - Persistent state for AI agent workflows
  14. For agentic coding, one can say: - Unless you need Terra Ultra perf, it's always better to use a Luna model with higher effort setting (sam…
  15. Yeah this is significantly better than Codex's subagent implementation. Turns out code is good, and sol is good at writing code! Programmat…
  16. Okay so basically: 5.6 Sol Ultra for planning Sol Medium for coding the plan and general coding tasks. Terra High for quick context subagen…
  17. We're launching Claude Tag today. Tag Claude into Slack and it works in channel with you. It’s proactive, multiplayer, with its own identit…
  18. Introducing Claude Tag, a new way for teams to work with Claude. In Slack, Claude joins as a team member with access to the channels and to…
  19. This is the start of Claude Everywhere. It’s Claude Code under the hood so it’s just as good at writing code. 65% of our product team’s new…
  20. calesthio/OpenMontage: World's first open-source, agentic video production system. 12 pipelines, 52 tools, 500+ agent skills. Turn your AI…
  21. google-labs-code/stitch-skills: A library of Agent Skills designed to work with the Stitch MCP server. Each skill follows the Agent Skills…
  22. addyosmani/agent-skills: Production-grade engineering skills for AI coding agents.
  23. I think I need to ship my own installation CLI for my skills. It means I could: - Ship skills more customized to individual agents (/claude…
  24. asgeirtj/system_prompts_leaks: Extracted system prompts from Anthropic - Claude Fable 5, Opus 4.8, Claude Code, Claude Design. OpenAI - Cha…
  25. Congrats to @SpaceXAI on Grok 4.5 — trained on NVIDIA GB300 NVL72 systems and purpose-built for coding, agentic tasks, and knowledge work.…

Sources & citations

x21715
github727
rss1401
hackernews601
reddit251
show_hn400
yc-rfs160
gmail80
yc-launch30
total58125