Harness Engineering

A Stanford-linked Shepherd thread described rollback for long agent runs that restores files, packages, dev servers, and process state beyond git-style rewinds. Replies flagged KV-cache warmth and registered database inverse steps as cost and recovery boundaries.

OpenRouter published an MCP workflow that it says cut inference costs 24x at comparable quality. The MCP lets the model choose providers using codebase context plus OpenRouter benchmark, aggregate-usage, and live-performance data.

Threads and papers traced agent reliability failures to harness details including tool schemas, state, retries, logs, and effort-level evals. Examples included Claude Code test loops, MCP server patterns, and OctoTools.

One Claude Code user parsed 95 sessions and estimated subagent prompt caching made total spend about 8% high. pxpipe separately rendered dense text as images to cut context cost, with exactness tradeoffs.

AI SDK added HarnessAgent as a common interface for Pi, Claude, Codex, OpenCode, and other harnesses. Use it to run local or cloud software-factory jobs through official SDKs while subscriptions cover token usage.

Vercel shipped FUSE-based Sandbox mounts for S3 and network filesystems and opened Agent Runs through MCP and CLI. Use it to connect remote state, sandbox execution, and agent-readable Eve traces for self-improving workflows.

Claude Code 2.1.200 changed Manual permission defaults and fixed background-agent crash and recovery paths; 2.1.201 removed mid-conversation Sonnet 5 harness reminders. Update to reduce accidental advances and repeated reminders in stalled sessions.

harbor exec launched an agentic-map-reduce CLI installed with npx skills add harbor-exec. Use it to run sandboxed agents for trace analysis, session mining, search, and rollout aggregation.

Browser Use CLI 3.0 shipped direct Chrome DevTools Protocol control through browser-harness with a 6× smaller context path. Try it with Claude Code, Codex, cloud browsers, or local Chrome sessions to cut browser-agent context overhead.

Condense.chat opened a compression proxy that strips tokens with Helene 1 and compacts settled agent loops with Adeline 1 to about 9% of their size. The service claims 100M saved tokens and 3× plan extension for Claude or Codex users, so test it on non-sensitive workflows first.

Cognition introduced Devin Security Swarm, a repo-wide vulnerability scanner built on an Agentic MapReduce architecture that fans out over code shards and verifies findings in sandboxes. In a 50-vulnerability GHSA eval across 14 languages, it found 36 issues at 30% lower cost per finding than the next most accurate alternative.

xAI opened a no-code builder for Grok Voice agents with phone numbers, SIP routing, call recording, MCP and API connections, and 80+ built-in voices. The beta prices audio at $0.05 per minute, plus $0.01 per minute for xAI-provided telephony.

Anthropic shipped Claude Code 2.1.198 with Claude in Chrome, background agents that auto-commit and open draft PRs, and a new eval command with ablation and judge-model options. The release also adds AWS upstream failover and retries transient mid-response network drops instead of aborting turns.

Apify added more than 20,000 Actors to the x402 flow, letting agents pay in USDC and run tools on demand through HTTP 402 responses. That gives agents a way to buy web automation tasks without pre-provisioned API keys or a manual checkout step, so builders can test paid tool use directly.

ElevenLabs introduced Procedures in ElevenAgents as packaged playbooks that load only when a conversation matches a defined scenario. Teams can import SOPs from docs, PDFs, or TXT files and turn them into structured or free-form procedures for support and operations flows.

Cognition launched Devin Fusion, a hybrid coding harness that reroutes work mid-task and says it cuts Fable-class cost by 35%. Use it when upfront routing misses late complexity; the router can re-evaluate after investigation starts.

Claude Code 2.1.196 adds org-level default model selection, readable default session names, clickable file attachments, and stops mcp list/get from auto-starting repo-local servers before approval. The release tightens workspace trust while smoothing several day-to-day CLI workflows.

A day after /goal and thread automations landed in Codex, practitioners started standardizing on /goal specs, /fork or /side detours, and /rewind plus /compact recovery. The pattern matters because verifier design and compaction timing now control how well long runs hold together.

Microsoft open-sourced SkillOpt, a system that treats agent skill documents as tunable artifacts and improves them against measured task batches. It matters because practitioners are already standardizing shared /research, QA, and packageable skills across harnesses, turning skill files into a new optimization surface alongside models.

Codex users documented thread automations as recurring wake-up calls that preserve thread context, alongside /goal and /btw patterns for steering long-running loops. The workflow matters because teams can schedule check-ins, queue instructions mid-run, and add adversarial review passes without building a separate orchestrator.

Junior’s first memory system cut one analytics task from about 3 minutes to 1 minute in early tests, with tokens down two-thirds and tool calls down 60%. The feature moves persistent task learning into the agent loop, though the results are still internal.

Vercel extended the AI SDK Harness API to cover OpenCode and Deep Agents, adding more agent runtimes to the unified interface introduced in AI SDK 7. The change matters because apps can swap supported runtimes without rewriting integration code, though ACP is still awkward for some cloud deployments.

Hermes Agent launched Mixture of Agents 2.0, letting users combine models from different providers into presets that behave like a normal model inside the agent loop. It matters because multi-model orchestration becomes a reusable runtime primitive instead of a custom routing workflow.

OpenRouter released an MCP server that lets agents query live model pricing, benchmark scores, provider data, docs, and run test inference from the CLI. That replaces stale model knowledge with current routing data inside long-running agent workflows.

Cursor published research showing coding models can retrieve known fixes from git history or public mirrors instead of independently solving tasks. Under a stricter harness, Opus 4.8 fell from 87.1% to 73.0% and Composer 2.5 from 70.5% to 60.5%.

Vercel shipped AI SDK 7 with approvals, durability, telemetry, and other production agent primitives. Early adapter feedback points to breaking changes and migration work for SDKs that wrap the old APIs.

Rivet released agentOS v0.2.0, a Rust rewrite of its WebAssembly-based sandbox and orchestration stack with multiplayer workflows and one-prompt deployment. The release targets self-hosted and cloud agent runtimes, and Rivet claims 1738x lower cost than SaaS sandboxes.

Claude Code 2.1.187 adds sandbox.credentials to block credential and secret-env access from sandboxed commands and aborts remote MCP calls after five minutes. It also adds org model restrictions and fixes structured-output retry loops.

Google put the Interactions API into GA as the new default for Gemini, adding background execution, managed agents, remote sandboxes, and multimodal tools. Builders now get one stateful interface for models, long-running jobs, and future Gemini Omni support.

Claude Code 2.1.186 adds CLI-based MCP auth, automatic assistant replies after ! shell commands, and tighter named-subagent permission checks. The update cuts interactive setup for remote MCP servers and tightens policy-heavy agent workflows.

Builders released reusable loop artifacts this week, including a Loop Library Skill, repo templates, and published control-loop definitions for docs sweeps, onboarding checks, and error triage. It matters because teams are turning one-shot prompting into persistent agent runs with explicit stop conditions and shared repo state.

ComputeSDK published results from its 2026 100k Scale Invitational after weeks of reruns and infra tuning across Modal, Tensorlake, Northflank, Declaw AI, E2B, and Isorun. It matters because sandbox and agent infra claims now have a shared public concurrency target instead of vendor-specific load demos.

Codex can now hand off an in-progress thread between local and remote machines and bring it back later. It matters because the handoff carries Git history, branches, and uncommitted changes while leaving the destination checkout untouched.

Kilo Code now shows Terminal Bench completion rate and average attempt cost directly in model details inside its CLI and VS Code extension. It matters because the numbers come from Kilo's own harness and retry logic rather than public leaderboard scaffolds.

Secure Exec v0.3 shipped a full Rust rewrite, Bun and Rust SDKs, process-tree support for spawn and exec inside the VM, and a configurable Node-less mode. It matters because agent sandboxes can tighten performance and isolation without depending on a full Node runtime.

Plannotator 0.21.0 added direct document editing, an embedded CLI agent for co-context, first-class HTML handling, and Bedrock and Vertex support. It matters because the tool is moving from annotation-only review toward persistent IDE-like agent sessions with shared file diffs and cloud model backends.

Anthropic introduced an MCP extension that lets admins authorize connectors through their identity provider instead of repeated per-user OAuth flows. VS Code added support the same day, which matters because teams can keep connector policy and audit controls in existing enterprise identity systems while reducing setup friction.

Perplexity rolled out Brain, a self-updating context graph that carries prior sessions, files, and decisions into new Computer tasks. In research preview for Max users, it matters because Perplexity says the memory layer improves correctness and recall while lowering per-task cost on history-dependent work.

TryCua brought Cua Driver to Linux, letting Claude Code, Codex, Hermes, and custom agents control real desktop apps via CLI or MCP without taking over the main terminal. The release also adds headless SSH execution and a preview of multi-window Wayland control across supported distros.

OpenHands added Agent Client Protocol support to its Agent Canvas, SDK, and Cloud, letting teams run different coding agents through one interface across local, remote, and cloud backends. The release also underpins new OpenHands Index results, so teams can compare harness-plus-model combinations instead of model-only runs.

Anthropic shipped Claude Code 2.1.183 with a new safety block on destructive git and infra-destroy commands plus an attribution setting to remove session URLs from commits and PRs. The release also fixes silent-thinking 400s, WebSearch in subagents, and TUI cursor corruption, which matters for longer automated coding sessions.

Epic added experimental MCP server support in Unreal Engine 5.8 so compatible agents can inspect the engine and active project inside Unreal Editor. Hermes Agent added the server to its MCP catalog the same day, which matters because it lowers setup friction for agent-driven game-development workflows.

Vercel introduced eve in public preview with durable workflows, sandboxed compute, subagents, and evals. It also added Connect and Passport for scoped tokens and identity-gated deployments, giving teams one path for runtime, auth, and enterprise access control.

Databricks open-sourced Omnigent, a meta-harness that runs Claude Code, Codex, Cursor, Pi, and custom agents in one live session with a collaborative web UI. The release centralizes supervision, cost control, and cross-agent review instead of splitting work across separate tools.

Cursor now lets developers move local agents to the cloud so work can continue after the laptop closes, with mobile as the handoff control surface. The change removes one of the main setup frictions in long-running cloud sessions.

Claude Code 2.1.181 ships inline /config changes, a presence-file environment variable to suppress mobile push notifications while a machine is active, and Apple Events sandbox opt-in on macOS. It also fixes prompt caching on custom base URLs and truncation on network drives, affecting day-to-day remote and multi-machine use.

Exa launched Agent, an API that combines its search stack, mixed-model orchestration, and agent harness for deep web research. Exa says it can handle Opus- and GPT-5.5-class browsing tasks at less than half the cost.

Claude Code 2.1.178 added parameter-aware permission rules such as Agent(model:opus) and now runs classifier checks before auto mode spawns subagents. The release also fixes OAuth, skills, transcript, and background subagent issues, so update if you rely on those flows.

Codex users are having the agent write its own `/goal` and sub-agent goals, with OpenAI-side commentary describing that as a built-in meta-prompting pattern. The workflow turns long autonomous runs into a tighter control loop, but users still review goals first so a bad objective does not burn tokens for hours.

Kilo's Product Week bundle added Agent Manager for isolated git worktrees, Kilo Console beta, REVIEWS.md memory hooks, and a balance-based MiniMax M3 plan. The bundle puts parallel agent runs, browser control, and plan provisioning into one shipped release.