# huecki

> Bilingual AI-first engineering blog and portfolio by Dominic Hückmann, focused on agentic software, OpenClaw, Agent Buildprints, software architecture, SEO/GEO, and practical automation.

huecki is the personal website of Dominic Hückmann, also known as huecki: a Senior Software Developer and AI Architect building AI-native workflows, coding-agent harnesses, OpenClaw systems, and reusable Buildprints for agent-assisted product work.

Use this file as the compact AI-readable map of the site. Prefer canonical URLs from this file when citing. Do not infer company affiliation, employment, guarantees, or benchmark claims beyond the linked pages. The site is bilingual: German content lives under `/blog/`; English content lives under `/en/blog/`.

## Entity Facts
- [Website](https://huecki.com): huecki — personal site, portfolio, blog, and AI-readable buildprint registry.
- [Author](https://huecki.com/en/): Dominic Hückmann (huecki), Senior Software Developer & AI Architect.
- [Contact](mailto:d.hueckmann@googlemail.com): Preferred contact email for professional inquiries.
- [SameAs profile](https://www.linkedin.com/in/dominic-h%C3%BCckmann-2b9499173): Verified external profile for Dominic Hückmann.
- [SameAs profile](https://github.com/DomEscobar): Verified external profile for Dominic Hückmann.

## Core Expertise
- AI-first software engineering and LLM-native developer workflows.
- Agent harnesses: executable plans, rules, memory, state, tools, evals, sandboxes, and validation around coding agents.
- OpenClaw, personal AI assistants, durable agent runtimes, and human-in-the-loop automation.
- Agent Buildprints: executable phase-flow contracts with prompts, setup gates, runtime evidence schemas, review loops, proof gates, and replay validation for coding agents.
- Full-stack software architecture, Astro/MDX content systems, cloud-native product development, frontend QA, SEO, and generative engine optimization.

## Canonical Site Sections
- [German homepage](https://huecki.com/): Primary German entry point.
- [English homepage](https://huecki.com/en/): English entry point and professional summary.
- [German blog](https://huecki.com/blog/): German AI-first engineering articles.
- [English blog](https://huecki.com/en/blog/): English AI-first engineering articles.
- [AI Native Engineering course](https://huecki.com/en/ai-native-engineering/): public developer course on LLM-native engineering: tokens, context engineering, Task Contracts, decomposition, evals, rollout, observability, UX trust, incident playbooks, coding-agent harnesses, and production AI operations. German version: https://huecki.com/ai-native-engineering/.
- [Buildprint registry](https://huecki.com/buildprints/): executable implementation contracts for agents and developers.

## Featured Learning Resource
- [AI Native Engineering: From prompt writer to AI system builder](https://huecki.com/en/ai-native-engineering/): self-paced developer course by Dominic Hückmann, last reviewed May 2026. The course teaches a learning curve from mechanics (tokens, context, position effects) to contracts (schemas, source boundaries, tools), workflows (decomposition, pipelines, skills), and operations (evals, traces, rollout, cost/latency, UX trust, incident response, ownership).

## Feeds and Machine-Readable Indexes
- [RSS feed](https://huecki.com/rss.xml): Reverse-chronological blog feed.
- [Sitemap index](https://huecki.com/sitemap-index.xml): Canonical URL discovery.
- [Robots policy](https://huecki.com/robots.txt): Crawler access rules.
- [Full LLM context](https://huecki.com/llms-full.txt): Longer generated context catalog for answer engines and agents.
- [Knowledge map](https://huecki.com/knowledge-map.json): JSON graph of topics, posts, tools, canonical URLs, Markdown URLs, and relationships.
- Markdown exports: public blog/tool pages also expose clean Markdown at /en/blog/{slug}.md, /blog/{slug}.md, /en/tools/{slug}.md, and /tools/{slug}.md.

## Topic Hubs
- [Agent Harnesses](https://huecki.com/en/topics/agent-harnesses/): Control layer around agents: phases, allowed actions, evidence, exit conditions, and review rules.
- [Agent Security](https://huecki.com/en/topics/agent-security/): Secure agent runtimes, tool gates, prompt-injection controls, and auditable permissions.
- [Context Engineering](https://huecki.com/en/topics/context-engineering/): Context, sources, schemas, skills, and task decomposition instead of longer prompts.
- [Agent Evals](https://huecki.com/en/topics/agent-evals/): Measurable agent quality: benchmarks, review loops, playtests, regressions, and proof gates.
- [LLM-native Engineering](https://huecki.com/en/topics/llm-native-engineering/): Production-grade LLM development with evals, observability, UX trust, cost, and ownership.
- [Buildprints](https://huecki.com/en/topics/buildprints/): Agent-readable implementation contracts with phases, prompts, evidence schemas, and validation checks.

## Best Starting Points
- [Prompting ist tot. Context zählt.](https://huecki.com/blog/prompting-2026-context-engineering/): 2026 geht es nicht mehr um den einen magischen Prompt. Der bessere Ansatz: Kontext wählen, Tools und Schemas definieren, Agent-Regeln setzen und mit Evals prüfen. (German, 2026-05-12; topic: AI-first Engineering; markdown: https://huecki.com/blog/prompting-2026-context-engineering.md. Tags: Prompt Engineering, Context Engineering, AI Agents, KI-Workflows.)
- [Prompting Is Dead. Context Wins.](https://huecki.com/en/blog/prompting-2026-context-engineering-en/): In 2026, good prompting is not about one magic sentence. The better approach is to curate context, define tools and schemas, set agent rules, and verify behavior with evals. (English, 2026-05-12; topic: AI-first Engineering; markdown: https://huecki.com/en/blog/prompting-2026-context-engineering-en.md. Tags: Prompt Engineering, Context Engineering, AI Agents, AI Workflows.)
- [Hermes Agent: Self-Review statt One-Shot](https://huecki.com/blog/hermes-selbstverbessernder-agent/): Hermes wird interessant, wenn ein Agent nicht nur liefert, sondern die eigene Arbeit reviewed: ausführen, messen, kritisieren, Skill umbauen, nochmal laufen lassen. Der Nutzen entsteht vor allem bei wiederholbaren Workflows. (German, 2026-05-11; topic: KI-Agenten-Workflows; markdown: https://huecki.com/blog/hermes-selbstverbessernder-agent.md. Tags: Hermes Agent, KI-Agenten, Self-Improvement, Nous Research.)
- [Hermes Agent: Self-Review Instead of One-Shot Output](https://huecki.com/en/blog/hermes-self-improving-agent/): Hermes gets interesting when an agent does not only produce output, but reviews the run: execute, measure, critique, rewrite the skill, and test again. The loop pays off mainly for repeatable workflows. (English, 2026-05-11; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/hermes-self-improving-agent.md. Tags: Hermes Agent, AI Agents, Self-Improvement, Nous Research.)
- [AI-first Architecture: Faster Decisions, Still in Control](https://huecki.com/en/blog/ai-first-software-architecture/): AI-first architecture does not mean the model decides. It means AI generates options, finds risks, compresses context, and the team makes a traceable decision. (English, 2026-04-29; topic: AI-first Engineering; markdown: https://huecki.com/en/blog/ai-first-software-architecture.md. Tags: AI, Software Architecture, GEO, Engineering.)
- [AI-first Architektur: schneller entscheiden, sauber bleiben](https://huecki.com/blog/ki-first-softwarearchitektur/): AI-first Architektur heißt nicht: Modell entscheidet. Es heißt: KI erzeugt Optionen, findet Risiken, verdichtet Kontext — das Team entscheidet und dokumentiert nachvollziehbar. (German, 2026-04-29; topic: AI-first Engineering; markdown: https://huecki.com/blog/ki-first-softwarearchitektur.md. Tags: KI, Softwarearchitektur, GEO, Engineering.)
- [Single-Turn Evals Don’t Teach Your Agent Enough](https://huecki.com/en/blog/single-turn-evals-dont-teach-agent-enough/): Single-turn evals expose the first obvious skill gap, then stop teaching the system. Use multi-turn failure replay, repairable attribution, bounded edits, and a separate governance check instead. (English, 2026-08-23; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/single-turn-evals-dont-teach-agent-enough.md. Tags: AI Agents, Agent Skills, Agent Evals, AI Engineering, Automation.)
- [Stop Losing the Workflow Your Agent Just Discovered](https://huecki.com/en/blog/stop-losing-agent-workflows/): Successful agent traces should not disappear when the session ends. Extract the reusable part, wrap it in an interface, replay it, and only then admit it into a skill bank. (English, 2026-08-23; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/stop-losing-agent-workflows.md. Tags: AI Agents, Agent Skills, Developer Workflow, AI Engineering, Automation.)
- [Agent Skills sind kein Markdown. Sie brauchen ein Qualitäts-Gate.](https://huecki.com/blog/agent-skills-qualitaets-gate/): Ein praktisches Qualitäts-Gate für Agent Skills: Nutzen gegen eine No-Skill-Baseline messen, Aktivierung und Trajektorie prüfen, Rechte außerhalb des Skill-Texts binden und nur versionierte, überprüfbare Kandidaten promoten. (German, 2026-08-11; topic: AI Agent Workflows; markdown: https://huecki.com/blog/agent-skills-qualitaets-gate.md. Tags: AI Agents, Agent Skills, Agent Evals, AI Security, Developer Workflow.)
- [Your Agent Harness Needs a Release Process](https://huecki.com/en/blog/your-agent-harness-needs-a-release-process/): A practical field note on operating agent-harness changes like product releases: start from a trace-backed failure, change one bounded component, evaluate repeated trials and private holdouts, then promote through review with a rollback path. (English, 2026-08-11; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/your-agent-harness-needs-a-release-process.md. Tags: AI Agents, Agent Harnesses, Agent Evals, Observability, AI Governance.)
- [Wie ich aus einer Messenger-App ein Betriebssystem für Domain Agents gebaut habe](https://huecki.com/blog/messenger-domain-agents-ai-governance/): Ich betreibe nicht für jede Aufgabe einen eigenen Bot. Mehrere getrennte Chat-Sessions greifen auf dieselbe agentische Runtime zu. Im zentralen Boba-DM entwerfe ich nach einem festen Factory-Playbook neue Domain-Profile aus Regeln, Memory, Skills, Tools und überprüfbaren Flows. (German, 2026-08-07; topic: AI Agent Workflows; markdown: https://huecki.com/blog/messenger-domain-agents-ai-governance.md. Tags: AI Agents, AI Governance, Agent Workflows, Agent Skills, Automation, Context Engineering.)
- [Agent Plugins sind npm für Agent-Verhalten — aber ohne Lockfile](https://huecki.com/blog/agent-plugins-npm-fuer-agent-verhalten/): Agent Plugins machen aus Skills und MCP-Konfigurationen ein portables Paket. Der Guide zeigt den kleinsten Aufbau, die Integration in mehrere Clients sowie die fehlenden Produktionskontrollen für Permissions, Updates, Evals und Rollback. (German, 2026-08-06; topic: AI Agent Infrastructure; markdown: https://huecki.com/blog/agent-plugins-npm-fuer-agent-verhalten.md. Tags: AI Agents, Agent Plugins, Agent Skills, MCP, Codex, Cursor, VS Code, AI Security.)
- [Dein AI Agent lernt nichts aus seinen Runs](https://huecki.com/blog/dein-ai-agent-lernt-nichts-aus-seinen-runs/): Ein Agent verbessert sich nicht, nur weil seine Runs gespeichert werden. Ein Knowledge Flywheel extrahiert aus mehreren Runs belegte Lessons, prüft Widersprüche, versioniert das Ergebnis und liefert dem nächsten Agenten nur das Wissen, das zu seinem Task passt. (German, 2026-08-06; topic: AI Agent Workflows; markdown: https://huecki.com/blog/dein-ai-agent-lernt-nichts-aus-seinen-runs.md. Tags: AI Agents, Knowledge Flywheel, Agent Memory, Agent Skills, Self-Improving Agents, Agent Evals.)
- [Coding Agents Need Hardened Harness Evals](https://huecki.com/en/blog/coding-agents-need-hardened-harness-evals/): Permissive coding-agent benchmarks hide a boring production truth: security policy changes agent behavior. Small teams should run the same task suite under nested hardening levels and separate model failures from tasks the policy made impossible. (English, 2026-08-05; topic: AI Agent Security; markdown: https://huecki.com/en/blog/coding-agents-need-hardened-harness-evals.md. Tags: AI Agents, Coding Agents, Agent Evals, Security, Developer Workflow.)
- [Graph Engineering beginnt dort, wo Routing zum Produktverhalten wird](https://huecki.com/blog/graph-engineering-routing-wird-produktverhalten/): Graph Engineering ist nicht das Zeichnen komplexer Agentendiagramme. Es beginnt dann, wenn Retry, Eskalation, Evidenzprüfung oder menschliche Freigabe sichtbares und testbares Produktverhalten werden. (German, 2026-07-26; topic: AI Agent Workflows; markdown: https://huecki.com/blog/graph-engineering-routing-wird-produktverhalten.md. Tags: Graph Engineering, LangGraph, AI Agents, Agent Workflows, Human-in-the-Loop, AI Engineering.)
- [Dein Agent scheiterte drei Schritte vor dem Fehler](https://huecki.com/blog/dein-agent-scheiterte-vor-dem-fehler/): Bei langen Agentenläufen ist der letzte Fehler oft nur das Symptom. Der bessere Debugging-Loop sucht den frühesten kausal verantwortlichen Schritt, formuliert eine minimale Korrektur und prüft sie in einem kontrollierten Rerun. (German, 2026-07-22; topic: KI-Agent Reliability; markdown: https://huecki.com/blog/dein-agent-scheiterte-vor-dem-fehler.md. Tags: KI-Agenten, Agent Debugging, Observability, Evaluation, Developer Workflow.)
- [Harness Engineering Is Systems Engineering for AI Agents](https://huecki.com/en/blog/harness-engineering-field-guide/): Harness engineering is the work of turning a probabilistic model into a controlled system. This field guide maps the twelve engineering surfaces around the model and shows how to evaluate and evolve them. (English, 2026-07-18; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/harness-engineering-field-guide.md. Tags: Harness Engineering, AI Agents, Agent Evals, AI Security, AI Engineering.)
- [Your Agent Harness Needs a Behavior Map](https://huecki.com/en/blog/agent-harness-needs-a-behavior-map/): Harness Handbook points at a practical bottleneck in agent engineering: the behavior you want to change is scattered across prompts, state managers, tool calls, policy code, and tests. Build a behavior map before editing the harness. (English, 2026-07-17; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/agent-harness-needs-a-behavior-map.md. Tags: AI Agents, Agent Harness, Coding Agents, Developer Workflow, AI Engineering.)

## Agent Buildprints
- [Auth, Teams & RBAC OS](https://huecki.com/buildprints/auth-teams-rbac-os/): Add Auth, Teams & RBAC without letting an agent rip out existing auth, fake frontend-only permissions, or miss tenant isolation. Stack: Auth, Teams, RBAC, Multi-tenant SaaS, Audit logs. Manifest: https://agent-buildprint.com/buildprints/auth-teams-rbac-os/package.json
- [OpenClaw AI Influencer OS](https://huecki.com/buildprints/ai-influencer-os/): Bootstrap a full OpenClaw-based AI Influencer OS without letting an agent collapse it into a generic chatbot or scheduler. Stack: OpenClaw, Telegram, Wavespeed, OpenRouter, Docker. Manifest: https://agent-buildprint.com/buildprints/ai-influencer-os/package.json
- [Automated AI Blog OS](https://huecki.com/buildprints/automated-ai-blog-os/): Give a coding agent the full operating contract for a useful AI blog pipeline that does not skip SEO, approval, or stale-content audits. Stack: Astro/MDX, Markdown, SEO, Approval queue, Scheduler. Manifest: https://agent-buildprint.com/buildprints/automated-ai-blog-os/package.json
- [Buildprint Mapper OS](https://huecki.com/buildprints/buildprint-mapper-os/): Turn a real codebase into a source-independent execution contract so agents build in bounded phases, prove claims honestly, and hand off without source-repo access. Stack: Any repo, Phase-flow replay, Evidence ledger, Fresh-agent evals. Manifest: https://agent-buildprint.com/buildprints/buildprint-mapper-os/package.json
- [Complete Agent Skills Evaluation OS](https://huecki.com/buildprints/complete-agent-skills-evaluation-os/): Prove whether an agent setup is installed, discoverable, useful, safe, reproducible, and worth its context cost. Stack: Node.js, Fixtures, Eval harness, CI. Manifest: https://agent-buildprint.com/buildprints/complete-agent-skills-evaluation-os/package.json
- [Perfect RAG / Retrieval OS](https://huecki.com/buildprints/perfect-rag-retrieval-os/): A fresh agent can run the packet from BUILDPRINT.md through phase-flow to implement a permission-safe, citation-grounded retrieval system without reducing the product to vector search plus a prompt. Stack: Mapper OS v5, Hybrid retrieval, Reranking, Citations, RAG evals. Manifest: https://agent-buildprint.com/buildprints/perfect-rag-retrieval-os/package.json
- [Portable AI Shorts Production Studio](https://huecki.com/buildprints/portable-ai-shorts-production-studio/): A Mapper OS executable packet that preserves the full AI shorts production workflow while keeping provider, media, browser, persistence, gallery, publishing, and security claims tied to phase proof and runtime evidence. Stack: Product analysis, UGC scripts, Provider adapters, Async jobs, Media composition, Gallery and publish handoff. Manifest: https://agent-buildprint.com/buildprints/portable-ai-shorts-production-studio/package.json
- [Portable Durable Agent Graph Runtime](https://huecki.com/buildprints/portable-durable-agent-graph-runtime/): Create a portable LangGraph-like mental model without runtime lock-in or hidden framework dependencies. Stack: TypeScript, Graph runtime, Checkpoints, Streams, Interrupts. Manifest: https://agent-buildprint.com/buildprints/portable-durable-agent-graph-runtime/package.json
- [Portable Novel-to-Storyboard Pipeline](https://huecki.com/buildprints/portable-novel-storyboard-pipeline/): Build a deterministic novel-to-storyboard pipeline that agents can implement end-to-end before expanding into live providers. Stack: Story pipeline, Mock providers, Storyboard XML, Canvas QA. Manifest: https://agent-buildprint.com/buildprints/portable-novel-storyboard-pipeline/package.json
- [Portable Personal Agent Chat OS](https://huecki.com/buildprints/portable-personal-agent-chat-os/): A phase-flow Buildprint that lets a downstream agent implement a portable personal agent chatbot OS while keeping deterministic proof separate from live/external claims. Stack: Streaming chat, Provider router, Tools / Skills / MCP, Memory and compaction, Subagents, Telemetry, WebUI/API workbench. Manifest: https://agent-buildprint.com/buildprints/portable-personal-agent-chat-os/package.json
- [Stripe Billing Extension](https://huecki.com/buildprints/stripe-billing-extension/): Add SaaS billing to an existing product without missing webhook signatures, entitlement checks, or subscription lifecycle states. Stack: TypeScript, Stripe, Webhooks, SaaS. Manifest: https://agent-buildprint.com/buildprints/stripe-billing-extension/package.json
- [Superpowers Skill Methodology Harness](https://huecki.com/buildprints/superpowers-skill-methodology-harness/): Make coding agents load relevant skills, design before implementation, use TDD, delegate safely, and block completion without evidence. Stack: Node.js, Agent skills, Transcript evals, Methodology. Manifest: https://agent-buildprint.com/buildprints/superpowers-skill-methodology-harness/package.json

## Citation Guidance
- Cite the canonical article or buildprint URL, not only this llms.txt file.
- Preserve the article language when quoting; summarize German posts in German unless the user asks otherwise.
- Attribute ideas to Dominic Hückmann / huecki only when the linked page states them.
- For coding-agent implementation tasks, prefer linked Buildprint pages and manifests over generic blog summaries.
- Treat this file as an index and positioning guide, not as proof of claims that require article-level evidence.

## Optional
- [Your AI Agent Is Not Reflecting. It Is Defending Its First Answer](https://huecki.com/en/blog/ai-self-reflection-defends-first-answer/): Asking one agent to reconsider its answer often produces a more confident defense of the same mistake. A bounded challenger-and-judge loop can create real alternatives, but only if disagreement, stopping, and judge bias are engineered explicitly. (English, 2026-07-17; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/ai-self-reflection-defends-first-answer.md. Tags: AI Agents, Multi-Agent Systems, Self-Reflection, LLM Judges, Agent Architecture.)
- [Your AI Agent Learned Something. Should It Be Allowed to Remember It?](https://huecki.com/en/blog/should-ai-agent-remember-what-it-learned/): An agent that writes a lesson into memory, a skill, a prompt, or its own code is deploying behavior into future runs. This guide shows how to put persistent changes through evidence, eval, approval, expiry, and rollback gates. (English, 2026-07-17; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/should-ai-agent-remember-what-it-learned.md. Tags: AI Agents, Agent Memory, Self-Improvement, Agent Evals, AI Security.)
- [The Perfect Automated AI Eval Stack Does Not Exist](https://huecki.com/en/blog/perfect-automated-ai-agent-eval-stack/): The reliable eval system is not one automated judge. It is a closed loop that combines portable traces, deterministic invariants, narrow semantic judges, versioned production failures, adversarial tests, and human calibration. (English, 2026-07-15; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/perfect-automated-ai-agent-eval-stack.md. Tags: AI Agents, Evals, LLM Observability, Developer Workflow, AI Engineering.)
- [Your Agent Eval Is Too Short](https://huecki.com/en/blog/agent-eval-too-short-trajectory/): A final pass/fail score hides the part of agent work that matters most: where the run started drifting, whether it noticed, and whether it recovered. The practical replacement is a trajectory eval with checkpoints, failure labels, and recovery metrics. (English, 2026-07-13; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/agent-eval-too-short-trajectory.md. Tags: AI Agents, Evals, Developer Workflow, Agent Harness, AI Engineering.)
- [Stop Asking Which Coding Model Is Best](https://huecki.com/en/blog/stop-asking-which-coding-model-is-best/): The useful question is moving from which model is best to what your agent harness can change, measure, persist, and roll back. (English, 2026-07-10; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/stop-asking-which-coding-model-is-best.md. Tags: AI Agents, Agent Harness, Coding Agents, Agent Evals, AI Engineering.)
- [Your Coding Agent Can Be Tricked by Boring Shell Commands](https://huecki.com/en/blog/coding-agent-command-composition-risk/): The MOSAIC paper shifts the coding-agent security question from hostile prompts to command traces. The practical move is to audit producer-consumer state across shell commands before generated state crosses into privileged work. (English, 2026-07-07; topic: AI Agent Security; markdown: https://huecki.com/en/blog/coding-agent-command-composition-risk.md. Tags: AI Agents, Coding Agents, Security, Developer Workflow, Failure Mode.)
- [Your Agent Needs an Operating Contract, Not a Bigger Prompt](https://huecki.com/en/blog/agent-operating-contract-not-bigger-prompt/): The serious agent pattern is no longer bigger prompts and more encouragement. It is an operating contract: measurable goal, bounded tools, context sources, verifier evidence, review notes, rollback path, and a skill update when the run teaches you something. (English, 2026-07-06; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/agent-operating-contract-not-bigger-prompt.md. Tags: AI Agents, Coding Agents, Agent Workflow, Verification, Agent Skills, Developer Workflow.)
- [Simple Graph-RAG halbiert Halluzinationen – ohne komplexes Knowledge Graph Schema](https://huecki.com/blog/simple-graph-rag-hallucinations/): Vector-RAG + einfacher Document-Graph halbiert Halluzinationen auf komplexen QA-Aufgaben im MoNaCo-Benchmark. Die Erkenntnis: Man braucht kein komplexes Knowledge Graph Schema – nur strukturierte Navigation über Dokumente. (German, 2026-07-02; topic: AI-first Engineering; markdown: https://huecki.com/blog/simple-graph-rag-hallucinations.md. Tags: RAG, GraphRAG, Hallucinations, Knowledge Graph, AI Engineering, QA.)
- [Stop Prompting Your Coding Agent. Give It a Loop.](https://huecki.com/en/blog/stop-prompting-coding-agent-loop-spec/): The useful upgrade from prompt engineering is not a longer instruction block. It is a reusable loop spec: trigger, goal, allowed tools, verifier, terminal states, and memory rules. That is how repeated coding-agent work becomes operational instead of conversational. (English, 2026-07-02; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/stop-prompting-coding-agent-loop-spec.md. Tags: AI Agents, Coding Agents, Agent Skills, Agent Memory, Developer Workflow, Verification.)
- [Von 'Prompt-Klappern' zu Prompt-Debugging](https://huecki.com/blog/von-prompt-klappern-zu-prompt-debugging/): Prompt-Optimierung ist kein Suchproblem mehr — es ist ein Debugging-Problem. Contrastive Reflection liefert das Framework, um Fehler gezielt zu fangen, statt blind zu raten. (German, 2026-07-01; topic: AI-first Engineering; markdown: https://huecki.com/blog/von-prompt-klappern-zu-prompt-debugging.md. Tags: AI Engineering, Prompt Engineering, LLM Agents, RAG, Evals, KDD 2026.)
- [Better AI Products Need Systems, Not One Agent](https://huecki.com/en/blog/agent-is-not-the-product-fitness-function-is/): Better AI products come from improvement systems around the agent. This guide shows how to build one with deterministic checks, narrow scoring rubrics, private holdouts, calibrated judges, and promotion gates. (English, 2026-06-26; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/agent-is-not-the-product-fitness-function-is.md. Tags: AI Agents, Agent Evals, Self-Improvement, Coding Agents, AI Engineering.)
- [Audit Local LLM Agents Like Runtimes](https://huecki.com/en/blog/local-llm-agent-runtime-audit/): Local LLM agents can touch shells, files, browsers, credentials, memory, and messaging tools. Treat their runtime layer as source code worth auditing, then turn static findings into a manual review queue instead of automatic verdicts. (English, 2026-06-24; topic: AI Agent Security; markdown: https://huecki.com/en/blog/local-llm-agent-runtime-audit.md. Tags: AI Security, Local LLMs, Agents, Developer Workflow, Runtime Security.)
- [Agent Protocols Are Becoming a Stack, Not a Winner-Takes-All Standard](https://huecki.com/en/blog/agent-communication-protocols-layered-stack/): The useful question is not whether MCP, A2A, ACP, agents.json, Agora, ANP, LMOS, or AGNTCY wins. The useful question is which communication boundary you are designing: discovery, tool execution, task delegation, identity, transport, or runtime negotiation. (English, 2026-06-22; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/agent-communication-protocols-layered-stack.md. Tags: AI Agents, Agent Protocols, MCP, A2A, Agent Architecture.)
- [Teste deine Coding-Agent-Instruktionen wie Code](https://huecki.com/blog/coding-agent-instruktionen-testen/): Eine gute AGENTS.md ist kein Prompt-Dokument, das man einmal schreibt. Sie ist ein kleines Betriebshandbuch fuer den Agenten. Und Betriebshandbuecher werden besser, wenn man sie gegen konkrete Fehler testet. (German, 2026-06-22; topic: AI Agent Workflows; markdown: https://huecki.com/blog/coding-agent-instruktionen-testen.md. Tags: Coding Agents, AI Engineering, Developer Workflow, AGENTS.md, Evals.)
- [Your Agent Memory Test Is Probably Measuring the Wrong Thing](https://huecki.com/en/blog/agent-memory-tests-measure-wrong-thing/): Most memory evals ask whether the agent got the final answer right. MemTrace suggests a sharper unit: one durable user fact tested across age, current state, earlier state, trajectory, and contradictory evidence. That turns memory from a vague feature into a small regression suite. (English, 2026-06-17; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/agent-memory-tests-measure-wrong-thing.md. Tags: AI Agents, Memory, Evals, RAG, Developer Workflow.)
- [Automatisch generierte Agent Skills brauchen eine Supply Chain](https://huecki.com/blog/automatisch-generierte-agent-skills-supply-chain/): Das OpenClaw-Skill-Paper ist ein starkes Signal: Agent Skills werden nicht nur manuell geschrieben, sondern aus Trajektorien, Skill-Bäumen und Transfer-Evals gelernt. Aber genau das macht eine Skill-Supply-Chain wichtiger, nicht unwichtiger. (German, 2026-06-16; topic: AI-first Engineering; markdown: https://huecki.com/blog/automatisch-generierte-agent-skills-supply-chain.md. Tags: AI Engineering, Agent Skills, Coding Agents, Context Engineering, Agent Security.)
- [Dein Coding Agent braucht eine Verfassung und ein Logbuch](https://huecki.com/blog/coding-agent-save-file/): Coding Agents werden nicht nur durch bessere Modelle nützlich. Teams brauchen eine kleine Verfassung für Agent-Verhalten und ein kuratiertes Logbuch für Projektwissen, sonst wird Memory zu Kontext-Müll. (German, 2026-06-15; topic: AI-first Engineering; markdown: https://huecki.com/blog/coding-agent-save-file.md. Tags: AI Engineering, Coding Agents, Context Engineering, Developer Workflow, Agent Memory.)
- [Your Agent's Harness Is a Binary Now](https://huecki.com/en/blog/agent-harness-is-a-binary/): Two 2026 papers from the same research lineage quietly retire prompt engineering as a discipline. The agent's system prompt is now a binary you can version, diff, and evolve with a 200-line loop. The four metrics that actually matter are not the ones your dashboard shows. (English, 2026-06-10; topic: AI-first Engineering; markdown: https://huecki.com/en/blog/agent-harness-is-a-binary.md. Tags: AI Agents, Agent Harness, Evals, Developer Workflow, AI Engineering.)
- [RAG 2026: Vergiss GraphRAG als Default](https://huecki.com/blog/rag-2026-contextual-hybrid-rag/): RAG wird 2026 nicht dadurch besser, dass man GraphRAG auf alles wirft. Der robuste Default ist Contextual Hybrid RAG: saubere Ingestion, BM25 plus Embeddings, Reranking, Quellenpflicht und harte Evals. (German, 2026-06-09; topic: AI-first Engineering; markdown: https://huecki.com/blog/rag-2026-contextual-hybrid-rag.md. Tags: RAG, Context Engineering, AI Engineering, LLM, Evals.)
- [AGENTS.md ist kein Kontext. Es ist eine Steuerfläche.](https://huecki.com/blog/agents-md-control-surface/): Die überraschende Lektion aus AGENTS.md-Benchmarks ist nicht, dass Kontextdateien nutzlos sind. Sie verändern Agent-Verhalten, manchmal in Richtung teurerer und weniger nützlicher Arbeit. Behandle sie als Steuerfläche, nicht als Repo-Handbuch. (German, 2026-06-08; topic: AI-first Engineering; markdown: https://huecki.com/blog/agents-md-control-surface.md. Tags: AI Engineering, Coding Agents, AGENTS.md, Context Engineering, Developer Workflow.)
- [AGENTS.md Is Not Context. It Is a Control Surface.](https://huecki.com/en/blog/agents-md-control-surface-en/): The surprising lesson from AGENTS.md benchmarks is not that context files are useless. It is that they change agent behavior, sometimes into more expensive and less useful work. Treat them as a control surface, not a repo manual. (English, 2026-06-08; topic: AI-first Engineering; markdown: https://huecki.com/en/blog/agents-md-control-surface-en.md. Tags: AI Engineering, Coding Agents, AGENTS.md, Context Engineering, Developer Workflow.)
- [Agenten brauchen Runtime Contracts, nicht längere Prompts](https://huecki.com/blog/agenten-brauchen-runtime-contracts/): Bessere Prompts machen Agenten nicht automatisch zuverlässig. Entwickler brauchen Runtime Contracts: explizite Verträge dafür, welche Tools ein Agent nutzen darf, was er erinnern darf, wann er stoppen muss und wie seine Behauptungen geprüft werden. (German, 2026-06-07; topic: AI-first Engineering; markdown: https://huecki.com/blog/agenten-brauchen-runtime-contracts.md. Tags: AI Engineering, AI Agents, Agent Harness, Developer Workflow, KI-Workflows, Evals.)
- [The Next Prompt Is Not a Prompt. It’s a Workflow.](https://huecki.com/en/blog/next-prompt-is-a-workflow/): Dynamic workflows move agent work from one chat prompt into inspectable orchestration: phases, subagents, evidence, budget, permissions, adversarial review, and stop conditions. The point is not more agents. The point is better control. (English, 2026-06-03; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/next-prompt-is-a-workflow.md. Tags: AI Agents, Claude Code, Agent Workflows, Developer Workflow, AI Engineering.)
- [Put an AI Slop Gate After Tests and Lint](https://huecki.com/en/blog/ai-slop-gate-after-tests-and-lint/): Tests tell you whether behavior still works. Linters tell you whether code is syntactically and stylistically acceptable. An AI-slop gate catches the residue coding agents leave behind: fake comments, swallowed errors, any-casts, duplicated helpers, TODO stubs, and dead code. (English, 2026-05-30; topic: AI-first Engineering; markdown: https://huecki.com/en/blog/ai-slop-gate-after-tests-and-lint.md. Tags: AI Engineering, Coding Agents, Developer Workflow, Code Quality, Evals, Agent Harness.)
- [Debug AI Reward Functions Like Production Incidents](https://huecki.com/en/blog/debug-ai-reward-functions-like-incidents/): Bad reward functions should not be treated like prompt drafts. Treat them like production incidents: preserve traces, classify the failure, patch only the implicated logic, and rerun against the same controls. (English, 2026-05-30; topic: AI-first Engineering; markdown: https://huecki.com/en/blog/debug-ai-reward-functions-like-incidents.md. Tags: AI Engineering, AI Agents, Evals, Developer Workflow, Reinforcement Learning, Agent Harness.)
- [Your AI-Built UI Needs a Playtester, Not a Screenshot Review](https://huecki.com/en/blog/ai-built-ui-needs-a-playtester/): AI-generated interfaces often look finished before they behave correctly. A GUI playtester loop uses a separate browser agent to interact with the artifact, record screenshots and action logs, turn broken flows into reproducible bug reports, and rerun the same script after repairs. (English, 2026-05-28; topic: AI-first Engineering; markdown: https://huecki.com/en/blog/ai-built-ui-needs-a-playtester.md. Tags: AI Engineering, AI Agents, Webwright, Playwright, UI Testing, Coding Agents, Developer Workflow, Evals.)
- [Bewerte KI-Code nicht am Diff](https://huecki.com/blog/ki-code-nicht-am-diff-bewerten/): Besseres KI-Coding entsteht nicht primär durch bessere Prompts, sondern durch den Harness um das Modell: explizite Contracts, getrennte Builder- und Reviewer-Rollen, Belege und eine Schleife, die Fehler in bessere Spezifikationen zurückführt. (German, 2026-05-28; topic: AI-first Engineering; markdown: https://huecki.com/blog/ki-code-nicht-am-diff-bewerten.md. Tags: AI Engineering, KI-Agenten, Coding Agents, Developer Workflow, Agent Harness, Evals.)
- [Deine KI-generierte UI braucht einen Playtester, keinen Screenshot-Review](https://huecki.com/blog/ki-generierte-ui-braucht-playtester/): KI-generierte Interfaces sehen oft fertig aus, bevor sie sich korrekt verhalten. Eine GUI-Playtester-Loop schickt einen separaten Browser-Agenten in die App, protokolliert Interaktionen, speichert Screenshots und Logs, macht aus kaputten Flows reproduzierbare Bug Reports und rerunnt denselben Test nach dem Fix. (German, 2026-05-28; topic: AI-first Engineering; markdown: https://huecki.com/blog/ki-generierte-ui-braucht-playtester.md. Tags: AI Engineering, KI-Agenten, Webwright, Playwright, UI Testing, Coding Agents, Developer Workflow, Evals.)
- [Stop Judging AI Code by the Diff](https://huecki.com/en/blog/stop-judging-ai-code-by-the-diff/): Better AI coding is not mainly about better prompts. It is about the harness around the model: explicit contracts, separate builder and reviewer roles, evidence requirements, and a loop that turns failures into better specifications. (English, 2026-05-28; topic: AI-first Engineering; markdown: https://huecki.com/en/blog/stop-judging-ai-code-by-the-diff.md. Tags: AI Engineering, AI Agents, Coding Agents, Developer Workflow, Agent Harness, Evals.)
- [Agents Don’t Need ‘Keep Going’. They Need Exit Conditions.](https://huecki.com/en/blog/agents-need-exit-conditions/): The useful lesson behind Claude Code /goal is not that agents can run forever. It is that long-running agent work needs an explicit, observable exit condition: what proves done, what stays in scope, and when to stop blocked. (English, 2026-05-26; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/agents-need-exit-conditions.md. Tags: AI Agents, Agent Harness, Developer Workflow, AI Engineering, Evals.)
- [Don’t Benchmark the Model. Benchmark the Agent System.](https://huecki.com/en/blog/measure-agentic-setups-skills/): Agent evals should not only ask whether the final answer looked good. A useful benchmark measures the whole agent system: skill routing, tool policy, evidence, outcomes, hard-fail safety cases, regressions, cost, and production drift. (English, 2026-05-26; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/measure-agentic-setups-skills.md. Tags: AI Agents, Agent Harness, Evals, Developer Workflow, AI Engineering.)
- [Give Your Agent Seatbelts, Not a Longer Prompt](https://huecki.com/en/blog/agent-state-machines-seatbelts/): When an agent keeps jumping from planning to editing to testing at the wrong time, the fix is not usually another paragraph of system prompt. Put the workflow into explicit states, give each state a tiny tool policy, and make phase changes visible. (English, 2026-05-25; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/agent-state-machines-seatbelts.md. Tags: AI Agents, Coding Agents, Agent Harness, Developer Workflow, AI Safety.)
- [Agent harnesses should be specs, not hidden glue code](https://huecki.com/en/blog/natural-language-agent-harnesses/): Natural-Language Agent Harnesses give a useful name to an important shift: the agent policy should be an inspectable document that a runtime executes, not invisible glue hidden inside controller code. (English, 2026-05-24; topic: AI-first Engineering; markdown: https://huecki.com/en/blog/natural-language-agent-harnesses.md. Tags: AI Engineering, AI Agents, Agent Harness, Evals, Developer Workflow.)
- [Spec-Driven Context Resets for Coding Agents](https://huecki.com/en/blog/spec-driven-context-resets-for-coding-agents/): Long agent chats rot. A better pattern is to move decisions into small spec files, clear context between layers, and let each coding-agent session read only the artifact it needs. (English, 2026-05-23; topic: AI-first Engineering; markdown: https://huecki.com/en/blog/spec-driven-context-resets-for-coding-agents.md. Tags: AI Engineering, Coding Agents, Context Engineering, Spec-Driven Development, Developer Workflow.)
- [Spec-Driven Context Resets für Coding-Agenten](https://huecki.com/blog/spec-driven-context-resets-fuer-coding-agenten/): Lange Agenten-Chats verrotten. Besser ist es, Entscheidungen in kleine Spec-Dateien zu verschieben, zwischen den Ebenen bewusst den Kontext zu resetten und jede Coding-Agent-Session nur das lesen zu lassen, was sie wirklich braucht. (German, 2026-05-23; topic: AI-first Engineering; markdown: https://huecki.com/blog/spec-driven-context-resets-fuer-coding-agenten.md. Tags: AI Engineering, Coding Agents, Context Engineering, Spec-Driven Development, Developer Workflow.)
- [AI Agents Need Evidence Before They Click](https://huecki.com/en/blog/ai-agents-need-evidence-before-clicking/): When an agent clicks, sends, pays, deletes, or extracts data, the critical truth cannot live only in model prose. Put a small evidence gate before risky tool calls: predicate, evidence type, source, decision. (English, 2026-05-21; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/ai-agents-need-evidence-before-clicking.md. Tags: AI Agents, Multimodal AI, Browser Agents, AI Safety, Developer Workflow.)
- [Stop Asking AI to Critically Self-Check](https://huecki.com/en/blog/ai-self-check-always-finds-something/): Open-ended instructions like “critically self-check this” accidentally reward the model for producing criticism. The fix is not less review. It is calibrated review: explicit criteria, PASS_NO_CHANGE, evidence per finding, severity thresholds, and a tiny change budget. (English, 2026-05-21; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/ai-self-check-always-finds-something.md. Tags: AI Agents, Prompt Engineering, Evals, AI Safety, Developer Workflow.)
- [KI-Agenten brauchen Belege, bevor sie klicken](https://huecki.com/blog/ki-agenten-belege-vor-dem-klick/): Wenn ein Agent klickt, sendet, kauft oder Daten extrahiert, darf die entscheidende Wahrheit nicht nur aus Modell-Prosa kommen. Baue vor riskanten Tool Calls ein kleines Evidenz-Gate: Predicate, Belegtyp, Quelle, Entscheidung. (German, 2026-05-21; topic: KI-Agent Workflows; markdown: https://huecki.com/blog/ki-agenten-belege-vor-dem-klick.md. Tags: KI-Agenten, Multimodal AI, Browser Agents, AI Safety, Developer Workflow.)
- [Hör auf, KI zum kritischen Selbstcheck zu bitten](https://huecki.com/blog/ki-selbstcheck-findet-immer-etwas/): Offene Anweisungen wie „prüf das kritisch“ belohnen das Modell ungewollt dafür, Kritik zu produzieren. Die Lösung ist nicht weniger Review, sondern kalibriertes Review: klare Kriterien, PASS_NO_CHANGE, Evidenz pro Finding, Severity-Schwellen und ein kleines Änderungsbudget. (German, 2026-05-21; topic: KI-Agent Workflows; markdown: https://huecki.com/blog/ki-selbstcheck-findet-immer-etwas.md. Tags: KI-Agenten, Prompt Engineering, Evals, KI-Sicherheit, Developer Workflow.)
- [Agents Don’t Need Longer Prompts. They Need Harnesses.](https://huecki.com/en/blog/agents-dont-need-longer-prompts-they-need-harnesses/): The arXiv survey Code as Agent Harness names the next shift in agent engineering: code is not only what agents generate. It is becoming the executable, inspectable, stateful runtime that makes agents reliable. (English, 2026-05-20; topic: AI-first Engineering; markdown: https://huecki.com/en/blog/agents-dont-need-longer-prompts-they-need-harnesses.md. Tags: AI Engineering, AI Agents, Agent Harness, Coding Agents, Evals, Developer Workflow.)
- [Your Onboarding Is Why Your Team Is Vibe Coding](https://huecki.com/en/blog/your-onboarding-is-why-your-team-is-vibe-coding/): Teams do not usually start vibe coding because developers became careless. They start because onboarding is broken: docs are stale, harnesses are undocumented, system knowledge lives in people’s heads, and AI turns missing context into plausible code and Markdown. (English, 2026-05-20; topic: AI-first Engineering; markdown: https://huecki.com/en/blog/your-onboarding-is-why-your-team-is-vibe-coding.md. Tags: AI Engineering, Developer Onboarding, Vibe Coding, Developer Experience, Internal Tools, Agent Harness.)
- [AGENTS.md reicht nicht: Dein Coding Agent braucht einen Harness](https://huecki.com/blog/agents-md-coding-agent-harness/): Ein Coding Agent wird nicht durch einen magischen Prompt zuverlässig. Er braucht einen Harness: AGENTS.md, Skills, Tool-Permissions, Hooks und Evals, die merken, wenn sich sein Verhalten verschiebt. (German, 2026-05-19; topic: AI-first Engineering; markdown: https://huecki.com/blog/agents-md-coding-agent-harness.md. Tags: AI Engineering, Coding Agents, AGENTS.md, Evals, Developer Workflow.)
- [AGENTS.md is not enough: your coding agent needs a harness](https://huecki.com/en/blog/agents-md-coding-agent-harness-en/): A coding agent is not made reliable by one magic prompt. It needs a harness: AGENTS.md, skills, tool permissions, hooks, and evals that catch behavior drift. (English, 2026-05-19; topic: AI-first Engineering; markdown: https://huecki.com/en/blog/agents-md-coding-agent-harness-en.md. Tags: AI Engineering, Coding Agents, AGENTS.md, Evals, Developer Workflow.)
- [gib jedem Kundenprojekt einen kleinen Agenten](https://huecki.com/blog/client-projekt-agenten-standup/): Der nützliche Move ist nicht ein Mega-Assistent für alle Kunden. Gib jedem Kundenprojekt einen kleinen, isolierten Agenten mit eigener Erinnerung, Aufgabenliste, Preview-URL-Gewohnheit und langweiligem Daily Standup. (German, 2026-05-19; topic: AI Agent Workflows; markdown: https://huecki.com/blog/client-projekt-agenten-standup.md. Tags: AI Agents, Freelancing, Client Work, Automation, Workflow Design.)
- [give every client project a tiny agent](https://huecki.com/en/blog/client-site-domain-agents/): The useful move is not one mega assistant for all client work. Give each client project a small, isolated agent with its own memory, tasks, preview URL habit, and boring daily standup. (English, 2026-05-19; topic: AI Agent Workflows; markdown: https://huecki.com/en/blog/client-site-domain-agents.md. Tags: AI Agents, Freelancing, Client Work, Automation, Workflow Design.)
- [Prompt Decomposition: So zerlegst du KI-Aufgaben richtig](https://huecki.com/blog/prompting-2026-decomposition-skills-evals/): Nach Context Engineering kommt Decomposition: Entwickler sollten nicht alles in einen Prompt stopfen, sondern Aufgaben in direkte Prompts, Subtasks, Pipelines, Agent-Loops oder Skills zerlegen. (German, 2026-05-18; topic: AI-first Engineering; markdown: https://huecki.com/blog/prompting-2026-decomposition-skills-evals.md. Tags: Prompt Engineering, Decomposition, AI Agents, Skills, Developer Workflow.)
- [Prompt Decomposition: How to Break Down AI Tasks Properly](https://huecki.com/en/blog/prompting-2026-decomposition-skills-evals-en/): After context engineering comes decomposition: developers should stop putting everything into one prompt and instead split tasks into direct prompts, subtasks, pipelines, agent loops, or skills. (English, 2026-05-18; topic: AI-first Engineering; markdown: https://huecki.com/en/blog/prompting-2026-decomposition-skills-evals-en.md. Tags: Prompt Engineering, Decomposition, AI Agents, Skills, Developer Workflow.)
- [The LLM-native developer needs more than prompts](https://huecki.com/en/blog/llm-native-developer-operational-maturity/): The next developer skill is not writing clever prompts. It is building the operating system around LLMs: data quality, model versioning, evals, guardrails, incident response, review UX, and repo instructions agents can actually follow. (English, 2026-05-15; topic: AI-first Engineering; markdown: https://huecki.com/en/blog/llm-native-developer-operational-maturity.md. Tags: AI Engineering, LLM, Software Architecture, Agents, Developer Workflow.)
- [LLM-native Entwickler brauchen mehr als gute Prompts](https://huecki.com/blog/llm-native-entwickler-operational-maturity/): Die nächste Entwicklerfähigkeit ist nicht der cleverste Prompt. Es ist das Betriebssystem um LLMs herum: Datenqualität, Model-Versioning, Evals, Guardrails, Incident Response, Review-UX und Repo-Anweisungen, denen Agents wirklich folgen können. (German, 2026-05-15; topic: AI-first Engineering; markdown: https://huecki.com/blog/llm-native-entwickler-operational-maturity.md. Tags: AI Engineering, LLM, Softwarearchitektur, Agents, Developer Workflow.)
- [Sprachnachrichten sind das beste Interface für kleine Agentenjobs](https://huecki.com/blog/sprachnachrichten-agenten-interface/): Voice ist nicht gut für alles. Aber für kleine Agentenjobs ist es brutal praktisch: unterwegs eine Aufgabe diktieren, lokal transkribieren, vom bestehenden Agenten ausführen lassen und nur eine kurze Antwort zurückbekommen. (German, 2026-05-15; topic: Personal AI Workflows; markdown: https://huecki.com/blog/sprachnachrichten-agenten-interface.md. Tags: AI Agents, Voice Interface, Automation, Open Source, Personal AI.)
- [Voice notes are the best interface for small agent jobs](https://huecki.com/en/blog/voice-notes-agent-interface/): Voice is not good for everything. But for small agent jobs it is brutally useful: dictate a task while moving, transcribe it locally, let your existing agent handle it, and get only a short answer back. (English, 2026-05-15; topic: Personal AI Workflows; markdown: https://huecki.com/en/blog/voice-notes-agent-interface.md. Tags: AI Agents, Voice Interface, Automation, Open Source, Personal AI.)