Development practices in the agentic erahttps://postcursors.dev/en-ushttps://postcursors.dev/en/til/rtk-find-90-pourcent-tokens-economises/https://postcursors.dev/en/til/rtk-find-90-pourcent-tokens-economises/TIL: inference-cost, workflow, toolingTue, 31 Mar 2026 00:00:00 GMThttps://postcursors.dev/en/til/sessions-ia-14-pourcent-produisent-commit/https://postcursors.dev/en/til/sessions-ia-14-pourcent-produisent-commit/TIL: workflow, agentic-workflowTue, 31 Mar 2026 00:00:00 GMThttps://postcursors.dev/en/dispatches/mythe-session-longue/https://postcursors.dev/en/dispatches/mythe-session-longue/Across 267 tracked sessions, 62% last under 10 minutes. Long sessions (>1h) account for only 15% of the total. The micro-session is the real productive pattern — not tool underuse. Fri, 27 Mar 2026 00:00:00 GMThttps://postcursors.dev/en/til/sessions-courtes-pattern-dominant/https://postcursors.dev/en/til/sessions-courtes-pattern-dominant/TIL: agentic-workflow, workflowTue, 24 Mar 2026 00:00:00 GMThttps://postcursors.dev/en/dispatches/sessions-ia-commencent-mal/https://postcursors.dev/en/dispatches/sessions-ia-commencent-mal/Across 128 auto-scored Claude Code sessions, 68 had a "wrong_approach" friction — 53%. The agent's first instinct is wrong half the time. This isn't a problem to fix, it's a pattern to integrate into the workflow. Fri, 20 Mar 2026 00:00:00 GMThttps://postcursors.dev/en/dispatches/kilo-code-opencode-equipe-feedback/https://postcursors.dev/en/dispatches/kilo-code-opencode-equipe-feedback/Kilo Code CLI is a fork of OpenCode with Kilo Gateway on top. In practice, the TUI is identical, custom commands work the same, and migration took 15 minutes. The real difference: free models via the gateway change the workflow economics. Tue, 17 Mar 2026 00:00:00 GMThttps://postcursors.dev/en/til/bash-read-edit-ratio-agent/https://postcursors.dev/en/til/bash-read-edit-ratio-agent/TIL: coding-agent, workflowSat, 14 Mar 2026 00:00:00 GMThttps://postcursors.dev/en/dispatches/trois-workflows-agentiques/https://postcursors.dev/en/dispatches/trois-workflows-agentiques/Three patterns that naturally emerged in a daily agentic workflow: automated morning briefing, versioned custom commands in Git, and parallel sessions in tmux splits. Each came from a real friction, not from theory. Fri, 13 Mar 2026 00:00:00 GMThttps://postcursors.dev/en/guides/agentic-workflow-principles/https://postcursors.dev/en/guides/agentic-workflow-principles/An agentic workflow isn't about "using AI to code faster" — it's a restructuring of how you break down, delegate, and verify work. The real gains come from three practices: decomposing tasks into verifiable units, actively managing context, and treating the agent like a competent junior who needs precise instructions. Thu, 19 Feb 2026 00:00:00 GMThttps://postcursors.dev/en/guides/cost-optimization-ai-coding/https://postcursors.dev/en/guides/cost-optimization-ai-coding/Route mechanical tasks to budget or free models (MiniMax 2.5 via Kilo Gateway, Gemini Flash), reserve premium models for reasoning. Result: mechanical tasks (~65% of volume) cost nothing, with no perceived quality loss. Thu, 19 Feb 2026 00:00:00 GMThttps://postcursors.dev/en/guides/prompt-engineering-for-coding-agents/https://postcursors.dev/en/guides/prompt-engineering-for-coding-agents/A good coding agent prompt has three components: what you want (precise task), how to validate it (verifiable exit condition), and what you don't want (explicit constraints). Most disappointing results come from instructions missing one of the three. Thu, 19 Feb 2026 00:00:00 GMThttps://postcursors.dev/en/guides/multi-agent-context-management/https://postcursors.dev/en/guides/multi-agent-context-management/The context window is a finite resource that degrades. The key is treating it as such: short sessions with verifiable deliverables, persistent context files between sessions, splitting into specialized agents for parallel tasks. Multi-agent workflows aren't added complexity — they're the way to work at scale without losing quality. Thu, 19 Feb 2026 00:00:00 GMThttps://postcursors.dev/en/guides/opencode-setup-neovim-tmux/https://postcursors.dev/en/guides/opencode-setup-neovim-tmux/OpenCode integrates naturally into a Neovim/tmux workflow: a dedicated tmux split, the agent in one pane, your editor in the other. The client/server architecture lets you start OpenCode server-side and connect via SSH with no extra configuration. Full setup in 30 minutes. Thu, 19 Feb 2026 00:00:00 GMThttps://postcursors.dev/en/dispatches/ambassadeur-ia-equipe-kilo-code-adoption/https://postcursors.dev/en/dispatches/ambassadeur-ia-equipe-kilo-code-adoption/The tool wasn't the problem, the model was. Kilo Code + Mistral looked perfect on paper. In practice, Mistral nearly killed adoption. Pivot to OVH AI Gateway + Gemini. It works now. Thu, 19 Feb 2026 00:00:00 GMThttps://postcursors.dev/en/dispatches/context-rot-organisation-pas-modele/https://postcursors.dev/en/dispatches/context-rot-organisation-pas-modele/The bigger the context window, the later context rot hits — it doesn't go away. The real solution is structuring your work into short sessions with verifiable deliverables. It's an organizational discipline, not a hardware problem to solve. Thu, 19 Feb 2026 00:00:00 GMThttps://postcursors.dev/en/dispatches/modeles-budget-vs-premium-coding/https://postcursors.dev/en/dispatches/modeles-budget-vs-premium-coding/On mechanical, repetitive PHP/TypeScript tasks, budget models (MiniMax 2.5 via Kilo Gateway, Gemini Flash) produce results indistinguishable from Claude Sonnet. On complex reasoning tasks, Claude maintains a real edge. The takeaway isn't "use the cheapest" — it's "use the right model for each type of task." Thu, 19 Feb 2026 00:00:00 GMThttps://postcursors.dev/en/dispatches/pourquoi-claude-code-vers-opencode/https://postcursors.dev/en/dispatches/pourquoi-claude-code-vers-opencode/Claude Code is an excellent starting point — zero config, immediate quality. But Anthropic lock-in prevents optimizing costs per task, and restricting the Claude Code subscription in third-party tools forced the migration. OpenCode then Kilo Code CLI (its fork with free gateway) solve both without sacrificing quality. Thu, 19 Feb 2026 00:00:00 GMThttps://postcursors.dev/en/til/modeles-budget-php-boilerplate/https://postcursors.dev/en/til/modeles-budget-php-boilerplate/TIL: kilo-code, inference-cost, phpThu, 19 Feb 2026 00:00:00 GMThttps://postcursors.dev/en/til/opencode-custom-command-api-platform/https://postcursors.dev/en/til/opencode-custom-command-api-platform/TIL: opencode, api-platform, phpThu, 19 Feb 2026 00:00:00 GMThttps://postcursors.dev/en/til/agents-md-par-sous-dossier/https://postcursors.dev/en/til/agents-md-par-sous-dossier/TIL: workflow, contextThu, 19 Feb 2026 00:00:00 GMThttps://postcursors.dev/en/til/opencode-parallel-sessions/https://postcursors.dev/en/til/opencode-parallel-sessions/TIL: opencode, neovim, workflowThu, 19 Feb 2026 00:00:00 GMT