Static archive of workflow research and patterns for AI-assisted development. 2026-04-07T00:00:00Z https://stayfresh.dev/ https://stayfresh.dev/workflows/claude-code-skills-stack/ 2026-04-07T00:00:00Z

A practical three-layer stack for Claude Code: decision, context, and execution.

https://stayfresh.dev/workflows/modular-workflow-stack/ 2026-04-07T00:00:00Z

Right tool, right job, right time: layered orchestration, typed agent delegation, human-in-the-loop gates, parallelism, and loops with real halting conditions.

https://stayfresh.dev/research/context-is-a-budget/ 2026-04-07T00:00:00Z

More tokens do not create more truth. Context should be spent where it changes the next decision.

https://stayfresh.dev/research/protocol-before-personality/ 2026-04-07T00:00:00Z

Shared artifacts and exit conditions beat charisma, cosplay, and fake seniority.

https://stayfresh.dev/research/specs-as-shared-reality/ 2026-04-07T00:00:00Z

Agents do not share understanding. They share binding artifacts, versions, and tests.

https://stayfresh.dev/research/what-is-prompting/ 2026-04-07T00:00:00Z

Prompting is a constrained pipeline where tokenization, max tokens, and control tokens form the operational boundaries that determine agent behavior.

https://stayfresh.dev/thesis/ 2026-03-29T00:00:00Z

The hard position behind the archive: reward surfaces, legible architecture, and tools that lower comprehension debt.

https://stayfresh.dev/workflows/anubis-github-pages/ 2026-03-29T00:00:00Z

How to put Anubis in front of a GitHub Pages site with a real reverse proxy and no proxy-loop nonsense.

https://stayfresh.dev/workflows/instant-project-sync/ 2026-03-19T00:00:00Z

Synchronize ticket queues against contract dates and repository evidence, with requirement reconstruction when docs are missing.

https://stayfresh.dev/workflows/psay-agent-notify/ 2026-03-19T00:00:00Z

Local voice and Notification Center loops for agent completion, handoff, and operator recall.

https://stayfresh.dev/workflows/project-ai-philosophy/ 2026-03-18T00:00:00Z

A written position for bounded, evidence-backed AI use in delivery.

https://stayfresh.dev/workflows/preference-toml/ 2026-03-18T00:00:00Z

Use RLHF-shaped semantics in a simple config DSL for agent evaluation loops.

https://stayfresh.dev/workflows/reward-rubric-dsl/ 2026-03-18T00:00:00Z

A small machine-readable format for scoring coding agent output.

https://stayfresh.dev/workflows/prompt-patterns/ 2026-03-18T00:00:00Z

Common prompt structures for reliable agent behavior.

https://stayfresh.dev/workflows/agent-psychology/ 2026-03-18T00:00:00Z

Understanding how agents reason and respond to instructions.

https://stayfresh.dev/workflows/ci-automation/ 2026-03-18T00:00:00Z

Integrating AI agents with continuous integration workflows.

https://stayfresh.dev/research/reward-engineering/ 2026-03-18T00:00:00Z

Why coding agents optimize the rubric more than the prompt.

https://stayfresh.dev/research/reward-hacking/ 2026-03-18T00:00:00Z

How poorly designed metrics produce plausible but unstable code.

https://stayfresh.dev/research/agents-md-effectiveness/ 2026-03-18T00:00:00Z

Evaluating repository-level context files for coding agents.

https://stayfresh.dev/research/formal-verification-agents/ 2026-03-18T00:00:00Z

Property-based testing and specification generation.

https://stayfresh.dev/research/enterprise-agent-design/ 2026-03-18T00:00:00Z

BCG framework for production-grade agents.

https://stayfresh.dev/research/persona-anchors/ 2026-03-18T00:00:00Z

Character-based instruction patterns for consistent agent behavior.