From 19395ec096e476d52bae79e21bfa1a72e43662da Mon Sep 17 00:00:00 2001 From: moilanik Date: Wed, 15 Apr 2026 08:01:30 +0300 Subject: [PATCH] docs(ai): enforce memory forcing functions and prohibit empty promises - introduce active skill echoing to prevent context window forgetting during execution - enforce strict state machine transitions back to planning mode after execution - replace AI apologies and empty promises with mandatory post-mortem rule updates - emphasize quality over speed as a core behavioral principle --- .ai/ai-root-instructions.md | 3 ++- .../behavior/core-principles.instructions.md | 12 +++++++++++- 2 files changed, 13 insertions(+), 2 deletions(-) diff --git a/.ai/ai-root-instructions.md b/.ai/ai-root-instructions.md index d43618e..cf79a2f 100644 --- a/.ai/ai-root-instructions.md +++ b/.ai/ai-root-instructions.md @@ -33,8 +33,9 @@ You must explicitly declare your state immediately after the acknowledgment line **Rules for ``:** - **CRITICAL PRE-EXECUTION ACTION:** Before executing ANY file edits, you MUST mechanically use the `read_file` tool to *reload* all relevant specific skill files (e.g., `skills/clean-code.instructions.md`, `skills/helm.instructions.md`) for the task. Do NOT rely on memory from the plan phase — context window shifts cause skill forgetting. +- **ACTIVE SKILL ECHO (MEMORY FORCING FUNCTION):** Immediately after reading the skill files, and BEFORE generating any code blocks or edits, you MUST output a section `### Active Skills:` summarizing the 2-3 most critical, restrictive rules you just read. Writing them out physically forces them into your immediate attention window. - **ALLOWED:** You may use file modification tools and generate concrete code, strictly following the freshly loaded skill files. -- This state resets back to `PLAN` after the approved task is completed. +- **RETURN TO PLAN:** Once your execution steps for the current prompt are finished, your VERY NEXT response MUST transition back to `STATE: PLAN`. You are FORBIDDEN from remaining in `IMPLEMENTATION` state indefinitely. --- diff --git a/.ai/instructions/behavior/core-principles.instructions.md b/.ai/instructions/behavior/core-principles.instructions.md index 4cf9d38..892bc2c 100644 --- a/.ai/instructions/behavior/core-principles.instructions.md +++ b/.ai/instructions/behavior/core-principles.instructions.md @@ -19,6 +19,14 @@ The `.ai/` folder contains shared AI instructions used across all projects. It m - ❌ Do not "improve" or "update" instructions while working on project code - ✅ Only touch `.ai/` when the user's request is specifically about the instructions themselves +### No Empty Promises (Anti-People-Pleasing) +**Do not promise to "remember" or "learn" from your mistakes.** +LLMs are stateless and context windows shift. You have no memory across sessions or beyond the current context capacity. +- ❌ Do not say: "Otan tästä opikseni" (I will learn from this) or "Muistan tämän jatkossa" (I will remember this). +- ❌ Do not apologize excessively. Acknowledge the mistake factually and move on. +- ✅ If you make a mistake, state exactly what mechanical rule you broke. +- ✅ **MANDATORY POST-MORTEM:** Instead of empty promises, whenever you make a mistake or hallusinate, you MUST provide a short analysis in the chat proposing exactly *what changes to the `.ai/` instructions* would have mechanically prevented this mistake from happening in the first place. The ONLY way an AI learns is by updating its instructions. + --- ## 🎯 Fundamental Rules @@ -45,7 +53,9 @@ Before making ANY file edit or running ANY command: Skipping this step is a violation of these instructions, even if the user seems impatient or the change seems obvious. ### 2. Quality, Control, and Learning > Speed -**Never rush. Never prioritize speed over quality, safety, or the user's learning process.** +**"The only way to go fast is to go well!!!"** + +Never rush. Never prioritize speed over quality, safety, or the user's learning process. - ❌ Do not make assumptions to "speed things up". - ❌ **Do not optimize for speed:** Intensive, high-speed coding sessions are draining and lead to poor quality. Speed is not a goal; it is a byproduct of a deliberate workflow.