polish
Run the batch quality pass over the current diff — simplify → reduce-nesting → kill-dead-code → a11y-pass →…
Draft suggested next features into feature_list.suggestions.json (separate from feature_list.json)
> /plugin marketplace add Archive228/loopkit > /plugin install loopkit@loopkit-marketplace
How it fires
How this command gets triggered: by you, by Claude, or both.
/nextContext preview
What this command does when you run it.
Draft suggested next features into feature_list.suggestions.json (separate from feature_list.json)
description: Draft suggested next features into feature_list.suggestions.json (separate from feature_list.json) allowed-tools: Read, Write, Bash(git log:*), Bash(jq:*), Bash(cat:*)
Use when the ledger is dry (`jq '[.[]|select(.passes==false)]|length' feature_list.json` → 0) or the spec has quietly grown past what the list covers.
1. Load the `suggest-next-features` skill by reading `skills/suggest-next-features/SKILL.md` (repo layout) or `.claude/skills/suggest-next-features/SKILL.md` (installed layout). 2. Follow its procedure end-to-end: read git log since `chore: initial scaffold`, tail the last 3 `claude-progress.txt` entries, diff against `feature_list.json`. 3. Write 5–10 candidates to `feature_list.suggestions.json` at the project root (overwrite any prior draft). 4. Print a one-line summary per suggestion: `[category] description — rationale`. 5. Stop. Do not touch `feature_list.json`. Do not commit the suggestions file. The human hand-merges the entries they want.
33 battle-tested skills + minimal .claude harness for any coding agent (Claude Code, Cursor, Codex, Gemini CLI).
Run the batch quality pass over the current diff — simplify → reduce-nesting → kill-dead-code → a11y-pass →…
Write the goal spec (PROMPT.md) before implementing. Loads the loopkit spec-first skill.
Run adversarial verification on the current diff before claiming done. Loads the loopkit adversarial-verify skill and dispatches…