{"slug": "i-built-an-approval-gate-for-ai-coding-agents-meet-friday-protocol", "title": "I Built an Approval Gate for AI Coding Agents — Meet Friday Protocol", "summary": "A developer has released Friday Protocol, an open-source skill that adds an explicit approval gate to AI coding agents, requiring a human to signal approval (\"OKE\") before an agent moves from planning to implementation. The workflow also enforces a verification loop that returns to implementation when tests fail rather than declaring a task complete, and is designed as a reusable layer rather than a replacement for existing coding agents. The developer describes it as an evolving project and a workflow-level safeguard, not a complete solution to AI agent safety.", "body_md": "AI coding agents can be incredibly productive.\n\nThey can inspect a codebase, propose a solution, modify files, run tests, and keep going with very little supervision.\n\nBut I kept thinking about one question:\n\n**What if the agent starts implementing the wrong plan before I have a chance to review it?**\n\nThe problem isn't necessarily that the AI cannot write code. Sometimes, the problem is that it can move from planning to execution too quickly.\n\nI wanted a workflow where an agent could still work autonomously, but not skip the moment when a human needs to make a decision.\n\nThat idea led me to build **Friday Protocol**.\n\nFriday Protocol is an open-source skill designed to introduce an approval-first workflow for AI coding agents.\n\nThe idea is straightforward:\n\nIn my workflow, `OKE` is the explicit approval signal.\n\nThe goal is not to make an AI agent less capable. It is to make the transition from planning to execution more deliberate and reviewable.\n\nI could simply add a sentence to a prompt:\n\n\"Ask me before making changes.\"\n\nBut I wanted something more structured than relying on a general instruction.\n\nFriday Protocol is built around explicit workflow gates, rather than treating approval as a polite suggestion.\n\nThe workflow is intended to prevent the agent from treating every request as permission to immediately execute its own interpretation of the task.\n\nIt also emphasizes verification. An agent should not declare a task complete merely because it has finished editing files.\n\nIf testing fails, the workflow should return to implementation and verification instead of pretending everything succeeded.\n\nOf course, the actual guarantees depend on the agent environment, how the skill is invoked, and how its instructions are followed. This is a workflow-level safeguard, not a claim that every possible unsafe action is technically impossible.\n\nHere is the simplified lifecycle:\n\n```\nTask Request\n     |\n     v\nInvestigation & Planning\n     |\n     v\nPresent Implementation Plan\n     |\n     v\nWait for Explicit Approval\n     |\n     v\nImplementation\n     |\n     v\nSecurity Checks & Testing\n     |\n     v\nDid Verification Pass?\n     |\n   +---+---+\n   |       |\n  No      Yes\n   |       |\n   v       v\nFix and   Report Results\nRetest    with Evidence\n```\n\nThe important part is the approval gate before implementation and the verification loop before completion.\n\nFriday Protocol is designed as a reusable skill, rather than a standalone application that must replace an existing coding agent.\n\nIt can be used as a workflow layer, with optional integration into my other project, MemoriaGraph, which explores contextual memory and decision support.\n\nI am also experimenting with these ideas in a broader security operations project. However, Friday Protocol itself focuses on the workflow and approval process, not on replacing a complete security operations platform.\n\nMy aim is to make the approach reusable across supported AI coding environments. Compatibility and behavior still need to be evaluated for each environment.\n\nOne thing I have learned is that adding more instructions is not automatically the same as creating a reliable workflow.\n\nA useful workflow needs observable checkpoints.\n\nFor example:\n\nThese questions are more useful than simply asking whether an agent claims to follow the rules.\n\nI want to keep improving Friday Protocol through practical testing, feedback, and clearer evidence of its behavior.\n\nIt is still an evolving open-source project, and I don't consider it a complete solution to AI agent safety.\n\nIf you work with AI coding agents, I would genuinely appreciate your feedback.\n\nRepository: [Friday Protocol on GitHub](https://github.com/lhermawan/friday-protocol)\n\nI'm particularly interested in hearing from developers who have experienced problems with agents moving too quickly, making changes before a plan is reviewed, or reporting completion without sufficient verification.\n\nA few questions I'd love to discuss:\n\nI'm building this from my own experience working with AI-assisted development, and I'd like to learn how other developers approach the same problem.\n\nThanks for reading. Feedback, criticism, and contributions are welcome.", "url": "https://wpnews.pro/news/i-built-an-approval-gate-for-ai-coding-agents-meet-friday-protocol", "canonical_source": "https://dev.to/lhermawan/i-built-an-approval-gate-for-ai-coding-agents-meet-friday-protocol-4646", "published_at": "2026-10-24 20:28:34+00:00", "updated_at": "2026-10-24 20:41:06.670630+00:00", "lang": "en", "topics": ["ai-agents", "ai-tools", "ai-safety", "developer-tools"], "entities": ["Friday Protocol", "MemoriaGraph", "GitHub"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/i-built-an-approval-gate-for-ai-coding-agents-meet-friday-protocol", "markdown": "https://wpnews.pro/news/i-built-an-approval-gate-for-ai-coding-agents-meet-friday-protocol.md", "text": "https://wpnews.pro/news/i-built-an-approval-gate-for-ai-coding-agents-meet-friday-protocol.txt", "jsonld": "https://wpnews.pro/news/i-built-an-approval-gate-for-ai-coding-agents-meet-friday-protocol.jsonld"}}