원클릭으로
db
Query the local SQLite database directly. Run SELECT queries, inspect schema, check data state.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
메뉴
Query the local SQLite database directly. Run SELECT queries, inspect schema, check data state.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Write Svelte 5 components and SvelteKit pages following the project's frontend conventions — dark theme, mobile-first, minimalist Tailwind, modern CSS.
Create a new Architecture Decision Record (ADR) in docs/decisions/. MUST be used whenever writing an ADR — never create ADR files manually. Picks the next sequential number, fills the template, and updates the ADR table in README.md.
Run E2E and unit tests. Supports smoke tests (quick, any DB), deterministic tests (seeded DB, full suite), unit tests, or individual test files. Also covers ad-hoc verification via Playwright MCP.
Audit the current branch for loose ends before merge or plan close — stale docs, untested defensive code, partial wiring, coverage regressions, leftover TODOs. Reports findings; does not fix. MUST be used as part of plan-mode step 4 ("Branch audit before closing"); also useful ad-hoc whenever a branch is asked to be merge-ready.
Stage and commit changes with an emoji conventional commit message. Reviews staged/unstaged diffs, suggests a message, and creates the commit.
Run linting and formatting checks on the project. Uses ESLint, Prettier, and svelte-check. Reports issues and offers to auto-fix.
| name | db |
| description | Query the local SQLite database directly. Run SELECT queries, inspect schema, check data state. |
| allowed-tools | Bash |
data/local.db (relative to project root)sqlite3 data/local.db "<SQL>"For readable output use: sqlite3 -column -header data/local.db "<SQL>"
category (id, slug, name, description, order, cutoff_s, cutoff_a, cutoff_b, cutoff_c, cutoff_d, cutoff_e, cutoff_f, created_at, updated_at)
tier_list_item (id, category_id→category, slug, name, description, score, order, created_at, updated_at)
UNIQUE(category_id, slug)
tag (slug PK, label, created_at, updated_at)
item_tag (item_id→tier_list_item, tag_slug→tag) PK(item_id, tag_slug)
page (slug PK, title, content, created_at, updated_at)
All foreign keys cascade on delete. Timestamps are ISO text columns with auto-update triggers.
-- Items in a category ordered by position
SELECT name, score, "order" FROM tier_list_item WHERE category_id = ? ORDER BY "order";
-- Items with their category name
SELECT i.name, i.score, c.name AS category FROM tier_list_item i JOIN category c ON c.id = i.category_id;
-- Tags for an item
SELECT t.label FROM item_tag it JOIN tag t ON t.slug = it.tag_slug WHERE it.item_id = ?;
-- Item counts per category
SELECT c.name, count(i.id) FROM category c LEFT JOIN tier_list_item i ON i.category_id = c.id GROUP BY c.id;
order column name — it is a reserved word: "order".If $ARGUMENTS is a SQL query, run it directly. If it is a natural language description, translate it to SQL first, then run it.
$ARGUMENTS