AgentStack Daily

AgentStack Daily EP059 leads with a current agent-stack release readout. Hermes Agent moves from `v0.14.0` to the `v0.15.x` line with a major kanban/orchestration expansion, faster launch and session search, promptware defense, Bitwarden Secrets Manager support, skill bundles, an Ink multi-session TUI, image provider additions, a Nous-approved MCP catalog, ntfy messaging, and follow-up fixes for dashboard loopback auth, Docker insecure-mode opt-in, MCP PATH resolution, `.md` media delivery, plugin packaging, and wheel/sdist manifests. Codex `rust-v0.135.0` adds richer `doctor` diagnostics, remote `/status` details, Vim text objects, named permission profiles, bundled zsh helper discovery, Python SDK sandbox presets, non-interactive install mode, and TUI/session reliability fixes. Claude Code `2.1.154` through `2.1.157` adds Opus 4.8, dynamic workflows, effort controls, automatic skill loading from `.claude/skills`, plugin scaffolding, `claude agents` worktree/session improvements, richer telemetry options, and many background-session, sandbox, MCP, image, terminal, and worktree fixes. The episode then covers Anthropic Opus 4.8, OpenAI workspace-agent controls, Chrome DevTools MCP, local code-flow tools, usage/quota telemetry, and programmable video rendering for agent-built media.

Show notes: https://tobyonfitnesstech.com/podcasts/episode-59/

AgentStack Daily EP058 leads with a new release gap: OpenClaw v2026.5.27, Codex rust-v0.134.0, and Claude Code 2.1.152/2.1.153 all landed after the last approved draft. OpenClaw strengthens prompt/content boundaries, no-auth exposure checks, Codex app-server recovery, metadata caching, durable channel delivery, provider coverage, embedding-provider support, Pixverse, DeepInfra catalog browsing, VLLM thinking params, Claude OAuth overlays, release checks, and CI proof paths. Codex adds local conversation-history search, a profile-first configuration model, better MCP setup with per-server environment targeting and OAuth for streamable HTTP, read-only MCP concurrency, richer hook and extension context, connector-schema preservation, and remote reliability fixes. Claude Code adds code-review fix mode, skill/tool restrictions, skill reload hooks, message-display hooks, plugin suggestion marketplaces, fallback-model continuity, background/workflow status improvements, MCP and remote fixes, then follows with LFS-skip plugin source options, update/doctor visibility, stricter subagent MCP policy handling, OAuth gateway credential fixes, macOS background-agent permission continuity, and many background-session repairs. The outside slate then covers MCP gateway governance, code-intelligence tools, shared agent memory and task state, mobile control bridges for local agents, local model routers and OpenAI-compatible serving, and DGX Spark / LM Studio as a private high-performance local model server.

Show notes: https://tobyonfitnesstech.com/podcasts/episode-58/

AgentStack Daily EP057 leads with OpenClaw v2026.5.22 and Claude Code 2.1.149/2.1.150, then broadens into source-verified AI news and practical GitHub-hosted projects. OpenClaw improves gateway startup behavior, plugin metadata reuse, meeting-notes capture contracts, chat-session pagination, Discord callback lifetimes, xAI/Grok search reuse, plugin SDK helpers, provider/media fallbacks, and reliability fixes. Claude Code 2.1.149 adds per-category usage accounting, keyboard-scrollable diff detail, task-list rendering, managed cloud MCP connector loading, and permission/sandbox/PowerShell/macOS safety fixes, while 2.1.150 is internal-only. Then the episode covers Gemini 3.5 Flash and Gemini API Managed Agents, Codex mobile/remote access and hybrid/on-prem enterprise movement, Anthropic's Stainless acquisition and Project Glasswing update, and a GitHub project radar focused on semantic code intelligence, MCP builders, model routing, local agents, security scanners, and multi-harness setup tools.

Show notes: https://tobyonfitnesstech.com/podcasts/episode-57/

AgentStack Daily EP056 opens with the release work that matters now: OpenClaw v2026.5.20 adds policy checks, safer secret-file handling, provider routing, cron and subagent recovery, voice context, and image-generation timeout fixes; Codex rust-v0.133.0 adds goals, remote-control readiness, permission-profile APIs, plugin discovery, and lifecycle hooks; Claude Code 2.1.148 follows a larger 2.1.147 update with pinned background sessions, code review, MCP pagination, enterprise policy enforcement, Windows behavior fixes, background approvals, and a Bash regression fix; Hermes Agent v0.14.0 adds the PyPI install path, a local OpenAI-compatible proxy for OAuth providers, SuperGrok auth, X search, Teams, lighter lazy installs, faster startup, browser speedups, handoff, LSP diagnostics, video generation, wider computer use, and more messaging surfaces. Then the episode moves through Codex Appshots, Secure MCP tunnels, Google Agent Executor, GKE Agent Sandbox and Agent Substrate, the Antigravity CLI migration, Microsoft MagenticLite, Google Data Agent Kit, Gemini API-key hardening, and Copilot Auto plus semantic issue search. Runtime target: 50 minutes.

Show notes: https://tobyonfitnesstech.com/podcasts/episode-56/

AgentStack Daily EP055 opens with the operator release readout: Codex CLI 0.132.0 adds first-class Python SDK authentication, simpler text turns, richer turn results, schema-constrained `codex exec resume`, faster TUI startup, auth-backed remote executor registration, image-fidelity preservation, goal-loop brakes, multi-session MCP replay fixes, remote websocket keepalives, and Windows install hardening. Claude Code CLI 2.1.145 adds `claude agents --json`, agent IDs in OpenTelemetry spans, GitHub repository and pull-request status in status-line JSON, richer plugin discovery before install, awaiting-input counts in terminal titles, hook payloads for background tasks and session crons, and several permission, MCP, terminal, review, plugin, and skill-loop fixes. Then the episode covers six concrete AgentStack topics: Google Gemini 3.5 Flash GA and Managed Agents, Chrome WebMCP, Google AI Studio's Workspace and Android build updates, Chrome DevTools for agents, and GitHub making GPT-5.3-Codex the base model for Copilot Business and Enterprise.

Show notes: https://tobyonfitnesstech.com/podcasts/episode-55/

AgentStack Daily EP054 opens on concrete release work: Claude Code CLI 2.1.144 stabilizes background and detached agent sessions, fixes a long startup hang when the API endpoint is unreachable, repairs MCP pagination and unsupported-image handling, adds background-session resume and a session-scoped model picker, and tightens read-before-edit and search-error behavior. Then five more builder-relevant moves: Cursor Composer 2.5, a Kimi K2.5-based coding model at roughly a tenth of frontier per-token cost; Anthropic acquiring Stainless and pulling SDK code generation in-house; Notion's Developer Platform turning the workspace into a hosted agent runtime with Workers and an External Agent API; the Vercel AI SDK rewriting its LangChain and LangGraph adapter; and Cloudflare Mesh putting zero-trust networking and identity under the agent lifecycle.

Show notes: https://tobyonfitnesstech.com/podcasts/episode-54/

AgentStack Daily EP053 opens with concrete release work: OpenClaw v2026.5.18 adds typed plugin tooling, faster gateway readiness, dialog-aware browser automation, runtime parity QA, realtime Android Talk Mode, safer media handling, stronger channel delivery, Codex app-server repairs, proxy TLS support, and operator-facing Mac app polish. OpenAI Codex CLI 0.131.0 adds richer TUI controls, unified mentions, plugin marketplace and sharing commands, daemon-managed remote control, configured remote environments, an `openai-codex` Python SDK, `codex doctor`, and tougher sandbox, auth, app-server, and state handling. Then the episode moves to GitHub's May 18 Copilot agent updates for remote CLI steering, cheaper model choices, and one-click Actions repair, followed by Anthropic's API update that gives Claude's web search tool richer SEC filing data for cited financial research workflows.

Show notes: https://tobyonfitnesstech.com/podcasts/episode-53/

This episode follows six concrete changes in the agent stack: Ollama pushing deeper into local coding-agent runtimes, LM Studio improving Apple Silicon vision inference and remote local servers, NVIDIA positioning DGX Spark as a serious local-agent machine, EXO showing where distributed local inference still needs hardening, xAI shipping Grok Build while redirecting older model slugs to Grok 4.3, and LiteLLM plus Envoy AI Gateway tightening the routing layer that sits between agents and models.

Show notes: https://tobyonfitnesstech.com/podcasts/episode-52/

AgentStack Daily EP051 opens with an agent-stack release readout: OpenClaw v2026.5.12 trims core installs, hardens Telegram, Codex, plugin, gateway, browser, and config paths, and improves reply delivery; Hermes Agent 2026.5.16 adds native Windows beta, PyPI installation, faster startup, a local OpenAI-compatible proxy, vision, video, browser, LSP, and verification upgrades; Claude Code 2.1.143 and 2.1.142 tighten plugin dependencies, background-session flags, PowerShell behavior, worktree isolation, MCP timeout handling, and agent-dashboard defaults. Then the episode turns to Google Cloud's Gemini Enterprise Agent Platform release notes for immutable agent revisions, traffic splitting, and Priority PayGo, and to Google's Interactions API breaking-change guide for the new `steps` timeline and `response_format` migration.

Show notes: https://tobyonfitnesstech.com/podcasts/episode-51/

This AgentStack Daily episode covers what is new in LLM and agent tooling: Hermes Agent v2026.5.7 adds durable boards, worker health checks, checkpoint pruning, gateway resume, no-agent cron, provider plugins, platform allowlists, and MCP fixes; Claude Code v2.1.141 through v2.1.129 adds the agent view, hook JSON updates, plugin and workload-identity controls, MCP repairs, and background-agent permission fixes; Google ADK documents pause-and-resume agents with persisted state; and GitHub exposes Copilot agent tasks through REST endpoints.

Show notes: https://tobyonfitnesstech.com/podcasts/episode-50/