← ClaudeAtlas

decision-loglisted

Log high-leverage decisions from the current work session to the Arete Decision Log. Extracts architecture, UX, scope, tooling, and philosophy decisions. Invoke with /decision-log or /adl.
AreteDriver/ai-skills · ★ 4 · AI & Automation · score 73
Install: claude install-skill AreteDriver/ai-skills
# Decision Log Skill Log high-leverage decisions from the current work session. ## Process 1. **Scan conversation** for decisions that: - Remove options or lock architecture - Choose between alternatives - Set scope boundaries - Select tooling/frameworks - Establish principles 2. **Generate ADL entry** for each: ```markdown ## ADL-YYYYMMDD-### **Project:** [from context] **Decision Class:** [ARCH|UX|SCOPE|TOOLING|PHIL] **Decision:** [one sentence] **Chosen Option:** [selected approach] **Rejected Options:** - [alternative 1] - [alternative 2] **Why:** - [reason 1] - [reason 2] **Tradeoffs Accepted:** - [what was given up] **Revisit If:** - [invalidation condition] ``` 3. **Route by Project visibility** (see Routing Rule below), then append to the chosen file. Before writing, two precheck commands: - `grep '^## ADL-YYYYMMDD-' <chosen-file>` to find today's highest sequence **for that file** and increment — sequence is **per-day per-file**, not shared across the two logs. - **Format-match check** (mandatory): `grep -A1 '^## ADL-' <chosen-file> | head -6` to confirm existing entries use the same field syntax (`**Project**:` vs `**Project:**` — both are seen in the wild; match what the file already uses). If the target file has zero ADL-shaped entries, STOP and surface to user: the path may have drifted. 4. **Report** what was logged, including which log it landed in and why. ## Routing Rule ARE