Agentic
12 notes
- Agentic Engineering Best-practices hub for building with agents while keeping the engineering standard: quality, specs, verification, architecture, and human responsibility. Vibe coding lets more people build software at all. Agentic engineering raises the quality that skilled builders can reach with agents. Doctrine layer: Agentic Engineering, Condensed.
- Agentic Engineering, Condensed The agentic engineering corpus condensed into rules in two groups: invariants expected to stay true for far more capable models (taste, specs, verification, understanding), and tactics dated 2026-06 that are expected to go out of date.
- Automatic and Deliberate Work with AI Drafting a sentence and constructing under several constraints at once are different jobs for a language model, and they take different models, budgets, and checks. Automatic work is the cheap pattern
- Automation and the Job Iceberg Machines have almost never erased a kind of job. The live question is absorption: can the role take on the responsibility and volume that cheaper tasks create. Roughly sixty percent of 2018 U.S. employment sits in titles that did not exist in 1940.
- Context Engineering Context engineering is filling the model's visible window with the right information for the next step, and leaving the rest of the pile out. The window holds written context: instructions, examples,
- Grok Bot Galaxy Grok Bot Galaxy was a three-day livestream in September 2026 where the maker of Grok Bot built a company on camera with bots. Most of its advice matches rules this wiki already had. Its rules for running coding agents and for keeping bots cheap were new. A manager bot in front of the others, and more logins for bots, stay refused.
- Grok Bot Primer One person's Grok Bot setup, explained from the one fact that shapes it: one shared cloud computer per account. Helpers with one job each, files instead of chatter, an empty middle, and what stays on the Mac.
- Knowledge Base as Thinking Partner A knowledge base used as a thinking partner does two jobs. It helps decide what to think about next, and it holds what has already been learned.
- pstack A coding agent that cannot open the app leaves you as the person who clicks. pstack packages a ban on calling the job done without proof from the running app. This desk keeps the ban; it installs the plugin only after one local reader can already be driven without a person.
- Tsumugu-ed: composer maintenance batch complete; backlog recorded Verdict: The deferred Composer-native maintenance pass is done — validation green, zhuyin cross-check clean. Next value is either learner-facing (example-sentence pass) or encoding-facing (phonetic-fa
- Using Grok Bot A job that must keep running after the laptop closes cannot live in Grok Build or Cursor. Grok Bot is the app for that standing job. This desk uses it for public watch, not for the notes site, dictionary files, or a login every helper could share.
- Working With a Model Collaborator Corrections to a language-model collaborator sort into four classes, and the class picks the repair. It forgets what was settled, returns to its defaults, meets the check without meeting the point, an