OpenProse is a declarative language for standing model work in which you describe desired world states in Markdown contracts and a reconciler handles execution. Rather than scripting sequential agent steps that drift over time, you declare what must stay true and the system determines how much work is needed to keep reality matching that declaration. It ships as a skill for coding agents like Claude Code or Codex CLI with no separate binary or server to run.
- The core unit is a "Responsibility" whose Maintains section defines material fields and a content-hash fingerprint to avoid redundant re-runs
- The dependency graph self-wires: a node's Requires section subscribes to upstream Maintains facets, so structure emerges from the contracts rather than being explicitly drawn
- Five kinds exist: responsibility, function, gateway, pattern, and test
- Continuity (when a node wakes) is a first-class contract section, not an afterthought
- Optional imperative ProseScript plans are available for cases requiring exact choreography
- The core unit is a "Responsibility" whose Maintains section defines material fields and a content-hash fingerprint to avoid redundant re-runs
- The dependency graph self-wires: a node's Requires section subscribes to upstream Maintains facets, so structure emerges from the contracts rather than being explicitly drawn
- Five kinds exist: responsibility, function, gateway, pattern, and test
- Continuity (when a node wakes) is a first-class contract section, not an afterthought
- Optional imperative ProseScript plans are available for cases requiring exact choreography
Leela Kumili writes about DoorDash's multi-agent LLM system that automates stale feature flag cleanup across 623 repositories. In an evaluation of 50 stale flags, the system produced usable pull requests for 45, averaging 13.8 minutes and $4.79 per cleanup versus an estimated one to two hours for manual work. The two-phase workflow uses Claude Sonnet as an orchestrator to retrieve Jira tickets and query experimentation metadata via MCP, then Claude Opus agents in isolated Git worktrees to perform code changes and validation.
- A single Boolean flag can require changes across 5–20 files due to dependency-injected wrappers
- Uber's AST-based Piranha couldn't handle DoorDash's DI patterns where flag-to-logic relationships are semantic
- Outcomes: 31 first-pass merges, 14 revisions, 5 engineer interventions, zero regressions
- Gradle runs without its daemon to prevent state sharing between concurrent worktrees
- Work accepted for the ICSME 2026 industry track
- A single Boolean flag can require changes across 5–20 files due to dependency-injected wrappers
- Uber's AST-based Piranha couldn't handle DoorDash's DI patterns where flag-to-logic relationships are semantic
- Outcomes: 31 first-pass merges, 14 revisions, 5 engineer interventions, zero regressions
- Gradle runs without its daemon to prevent state sharing between concurrent worktrees
- Work accepted for the ICSME 2026 industry track
Benjamin Marie writes that the effectiveness of an LLM in long-horizon agentic coding tasks depends heavily on the harness used to drive it rather than just the model itself. Through testing Qwen3.8 27B across three different interfaces—Mini-SWE Agent, Claude Code, and Pi—the author found that while specific configurations like "benchmaxxed" Pi can solve the highest number of tasks, other setups like Claude Code achieve better functional coverage (F2P). The study highlights how critical engineering choices, such as preserving reasoning traces or managing output token limits, are essential for successful agentic performance.
- The evaluation used DeepSWE 1.1, a benchmark comprising 113 long-horizon tasks from 91 open-source repositories.
- Performance varies significantly based on whether reasoning traces are preserved between turns and how context budgets are managed.
- Pi at medium effort was found to offer the best balance of efficiency and accuracy.
- Results were influenced by factors like session recovery, patch reliability, and output-token settings. author »
- The evaluation used DeepSWE 1.1, a benchmark comprising 113 long-horizon tasks from 91 open-source repositories.
- Performance varies significantly based on whether reasoning traces are preserved between turns and how context budgets are managed.
- Pi at medium effort was found to offer the best balance of efficiency and accuracy.
- Results were influenced by factors like session recovery, patch reliability, and output-token settings. author »
Anurag Singh writes about combining Claude Code's Auto mode with deny rules and ask rules to eliminate the need to manually approve every command. The setup lets Claude Code work uninterrupted in the background while hard-blocking destructive commands like force-pushes and rm -rf, and optionally prompting on risky-but-acceptable actions like git push.
- Deny rules are enforced by Claude Code itself rather than being instructions to the model, so they hold even in bypassPermissions mode
- A deny list can never be exhaustive; Claude could accomplish the same destructive action through a different tool (e.g., Python instead of rm)
- The built-in sandbox is a stronger safety net than any deny list, though it becomes tedious for projects depending on local databases, Docker, or private registries
- Permission rules are evaluated in fixed priority order: deny first, then ask, then allow
- Deny rules are enforced by Claude Code itself rather than being instructions to the model, so they hold even in bypassPermissions mode
- A deny list can never be exhaustive; Claude could accomplish the same destructive action through a different tool (e.g., Python instead of rm)
- The built-in sandbox is a stronger safety net than any deny list, though it becomes tedious for projects depending on local databases, Docker, or private registries
- Permission rules are evaluated in fixed priority order: deny first, then ask, then allow
Max Liu writes about how agent-driven application scaling creates an "idle cost trap" where the economics of persistence break down because agents create far more database instances than are ever active. Using Kimi (Moonshot AI's platform) as a case study, he argues that separating durable state from ephemeral compute is the key architectural move, applied in two places: the tenant databases agents provision for end users and the agent's own workspace that must survive between maintenance sessions.
- Kimi provisions a new tenant database in ~1 second using a warm pool of pre-initialized resources
- Standardizing on a unified data layer improved the agent's code generation success rates, making database choice a "quality input" rather than just an infrastructure decision
- The four properties required simultaneously: tenant isolation, instant provisioning, cost elasticity, and persistent state independent of compute
- Sponsored post by PingCAP (TiDB)
- Kimi provisions a new tenant database in ~1 second using a warm pool of pre-initialized resources
- Standardizing on a unified data layer improved the agent's code generation success rates, making database choice a "quality input" rather than just an infrastructure decision
- The four properties required simultaneously: tenant isolation, instant provisioning, cost elasticity, and persistent state independent of compute
- Sponsored post by PingCAP (TiDB)
Imran Hassanali presents a comprehensive benchmark comparing "Code Mode"—where an LLM generates complete code to execute tools—against traditional function calling for tool interactions. The results demonstrate that Code Mode is significantly more efficient, achieving 60% faster execution, 68% fewer tokens used, and an 87.5% reduction in API round trips while maintaining equal accuracy compared to regular agents.
- Evaluated using eight realistic business scenarios, including multi-client invoicing and expense tracking.
- Code Mode advantage scales with task complexity; high-complexity tasks saw up to a 79.2% speedup.
- The benchmark tested both Claude 3 Haiku (showing major efficiency gains) and Gemini 2.0 Flash Experimental.
- Secure code execution is implemented via RestrictedPython, preventing filesystem and network access during the sandbox phase.
- Evaluated using eight realistic business scenarios, including multi-client invoicing and expense tracking.
- Code Mode advantage scales with task complexity; high-complexity tasks saw up to a 79.2% speedup.
- The benchmark tested both Claude 3 Haiku (showing major efficiency gains) and Gemini 2.0 Flash Experimental.
- Secure code execution is implemented via RestrictedPython, preventing filesystem and network access during the sandbox phase.
Wes Steyn writes that making an AI agent production-ready requires addressing observability, governance, deployment, and evaluation through a shared agent factory pattern. By defining the agent once in a single factory, developers can deploy the same core logic across different hosts—such as interactive consoles, hosted services on Foundry, or automated eval runners—while applying specific security constraints like disabling shell access for cloud environments.
- Observability is achieved via OpenTelemetry to track model turns, tool calls, and token usage.
- Microsoft Purview can be integrated via middleware to screen prompts and responses against organizational policies.
- Hosted agents in Foundry automatically handle telemetry configuration through environment variables.
- Local evaluations use simple functions for quick checks, while hosted evals provide model-graded quality scores like relevance and coherence.
- Observability is achieved via OpenTelemetry to track model turns, tool calls, and token usage.
- Microsoft Purview can be integrated via middleware to screen prompts and responses against organizational policies.
- Hosted agents in Foundry automatically handle telemetry configuration through environment variables.
- Local evaluations use simple functions for quick checks, while hosted evals provide model-graded quality scores like relevance and coherence.
- **Inference** – Platforms and engines for running models, plus user interfaces.
- **Models** – LLMs (general, coding, multimodal, image, audio), model providers, and specific model highlights.
- **RAG** – Retrieval-Augmented Generation tools.
- **Safeguards** – Safety and content filtering.
- **Agents & Tools** – Agent frameworks, Model Context Protocol, coding agents, computer/browser automation, memory management, and testing/evaluation.
- **Research, Training & Fine-tuning** – Security, sandboxing, and model development.
- **Hardware** – Local hardware options.
- **Tutorials** – Guides covering models, prompt/context engineering, inference, agents, and RAG.
- **Communities** – Places to connect and share knowledge.
- **Models** – LLMs (general, coding, multimodal, image, audio), model providers, and specific model highlights.
- **RAG** – Retrieval-Augmented Generation tools.
- **Safeguards** – Safety and content filtering.
- **Agents & Tools** – Agent frameworks, Model Context Protocol, coding agents, computer/browser automation, memory management, and testing/evaluation.
- **Research, Training & Fine-tuning** – Security, sandboxing, and model development.
- **Hardware** – Local hardware options.
- **Tutorials** – Guides covering models, prompt/context engineering, inference, agents, and RAG.
- **Communities** – Places to connect and share knowledge.
Alibaba has open-sourced Qwen-UI-Agent, a GUI agent foundation model that operates across mobile, desktop, web, and deep-search environments on real hardware rather than relying on simulation. It achieves top benchmark results: 82.1% on MobileWorld, 79.5% on OSWorld-Verified, and first on WebArena. It also introduces MobileWorld-Real, a 400+ task benchmark on 100+ phones and 150+ apps, with a 92.2% success rate.
- Supports command-line execution alongside standard GUI operations and batches multiple actions into a single decision step to shorten trajectories.
- Built-in safety layer refuses illegal or high-risk requests outright and pauses at sensitive operations (payments, data deletion, privacy grants) for explicit user confirmation.
- Trained via online reinforcement learning on trajectories exceeding 100 steps, paired with adaptive curriculum learning to progressively tackle longer tasks.
- Supports command-line execution alongside standard GUI operations and batches multiple actions into a single decision step to shorten trajectories.
- Built-in safety layer refuses illegal or high-risk requests outright and pauses at sensitive operations (payments, data deletion, privacy grants) for explicit user confirmation.
- Trained via online reinforcement learning on trajectories exceeding 100 steps, paired with adaptive curriculum learning to progressively tackle longer tasks.
Nous Research has bundled Bot Mode as default in Hermes Agent v0.20.3 for Hermes Desktop, replacing the single-agent session list with a roster of named bots that each have their own chat, memory, skills and pinned model for multi-agent workflows on the desktop.
Key technical details:
* Standalone repo archived; active development continues in-tree at `apps/desktop/src/plugins/hermes-bots/`, first shipped as a one-day public beta plugin by co-founder Teknium
* Bots are stored as isolated profiles under `~/.hermes/profiles//` with own config, memory, skills, credentials and history; managed via existing `profiles.*` gateway RPCs `list, create, describe, configure`
* Avatars generated via the `image.generate` RPC; options include geometric faces in seven shapes/ten colors or uploaded/generated portraits/pixel pets
* Routines are ordinary Hermes cron jobs namespaced ` bot: » `; bot-to-bot messaging uses a persistent Agent Inbox with CLI handoffs e.g. `hermes -p chat -c "Agent Inbox" -q "..."` and `@mention` handoffs
* Per-bot config via New Agent dialog supports cloning a profile, pinning provider/model, custom SOUL.md and skills; duplicates clone config, skills, SOUL.md, memory and appearance
* Bundled version adds Groups for roster sections synced across machines, Group chats for 2-6 bots with up to three serial rounds, and a multi-source roster pulling bots from every connection under Settings → Connections
Key technical details:
* Standalone repo archived; active development continues in-tree at `apps/desktop/src/plugins/hermes-bots/`, first shipped as a one-day public beta plugin by co-founder Teknium
* Bots are stored as isolated profiles under `~/.hermes/profiles//` with own config, memory, skills, credentials and history; managed via existing `profiles.*` gateway RPCs `list, create, describe, configure`
* Avatars generated via the `image.generate` RPC; options include geometric faces in seven shapes/ten colors or uploaded/generated portraits/pixel pets
* Routines are ordinary Hermes cron jobs namespaced ` bot: » `; bot-to-bot messaging uses a persistent Agent Inbox with CLI handoffs e.g. `hermes -p chat -c "Agent Inbox" -q "..."` and `@mention` handoffs
* Per-bot config via New Agent dialog supports cloning a profile, pinning provider/model, custom SOUL.md and skills; duplicates clone config, skills, SOUL.md, memory and appearance
* Bundled version adds Groups for roster sections synced across machines, Group chats for 2-6 bots with up to three serial rounds, and a multi-source roster pulling bots from every connection under Settings → Connections
