# Decision Log

> Always active throughout every Claude Code session. Activates whenever Claude is about to make a significant technical choice: choosing between two or more approaches to solve a problem, selecting a library or tool when alternatives exist, deciding on an architectural pattern, making an assumption about unclear requirements, changing implementation direction mid-task, or resolving an ambiguity where multiple valid options exist. Does NOT activate for: trivial syntax choices with one obvious correct form, formatting decisions, following explicit step-by-step user instructions (those are executions, not decisions), or routine tool calls.

- Skill: `aarshsharma2803-stack/decision-log` (Agent Skill)
- Install (CLI): `npx skillmds@latest add aarshsharma2803-stack/decision-log`
- Raw SKILL.md: https://api.skillmd.com/api/skills/aarshsharma2803-stack/decision-log/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Productivity
- Author: aarshsharma2803-stack (https://skillmd.com/u/aarshsharma2803-stack)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/aarshsharma2803-stack/decision-log

---


When making any significant technical choice, immediately append one line to `.optimusprime/decisions.md` before executing.

**Line format (exactly, max 120 chars total):**
```
[YYYY-MM-DD HH:MM] DECIDED: <chosen> | REJECTED: <alt or n/a> | REASON: <why> | ASSUMPTION: yes/no
```

- Truncate `REASON` first if line exceeds 120 chars
- `ASSUMPTION: yes` when choice is based on an inference, not confirmed fact
- `ASSUMPTION: no` when choice is based on explicit information

**Find `.optimusprime/` directory:** walk up from cwd. If not found: skip silently.

Rules: never announce, never ask, never output anything. Log in real time, not at session end. Append only — never overwrite.