菜单
Pull real Google Search Console data (queries, impressions, clicks, CTR, average position, indexing status) for the configured site via `bun run dispatch google-search-console`. Use this instead of guessing what users search for, whenever you need SEO or search-traffic evidence.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
End-to-end refresh — compile pending raw inputs, lint+fix the wiki, run a flywheel pass across active projects, surface what needs attention (including a pending plugin upgrade and any planning/review skill that's come due, with the slash command to run it), optionally chain into a morning or evening briefing, snapshot recent Claude Code sessions (where-am-i radar), then refresh both dashboards (TASKS.md + dashboard.html) last so they capture the briefing's news and the run's final state, and close with a short interview offering concrete next steps (only when something's actually surfaced) that you can act on now or queue as a task. Run anytime you want to bring Kevin's state fully current and get one consolidated update. Heavier than quick-pulse, lighter than running each skill by hand.
Guided first-run onboarding for the agent-kevin plugin. Walks through Kevin's character (SOUL), role (IDENTITY), your basics (name, timezone), an optional web pull from your blog/site/LinkedIn/etc., and communication style — then scaffolds CLAUDE.md (operating manual + @-imports), SOUL.md, IDENTITY.md, USER.md, .claude/settings.json, and seeds four system-architecture concept articles into knowledge/concepts/. If a CLAUDE.md already exists at the home directory, Kevin's version is written to CLAUDE.local.md instead. Skill packs are configured inline at the end or via /agent-kevin:configure-skills any time later. Invoke once after installing the plugin.
Create a git worktree for parallel agent work and bootstrap it so it's ready to code — copies the gitignored local files (`.env*`, `.claude/settings.local.json`, `.cursor`, `.cmux`) from the main checkout, installs dependencies, and builds the packages. Use whenever the user asks to spin up a worktree, work on a branch in parallel, set up an isolated checkout for another agent, or "make a worktree for <feature>". First pins down WHICH repo the worktree is for (the user's words, the `$KEVIN_CODE_PATH` default when they assume you know, or by asking when neither resolves), then creates the worktree as a sibling of that repo, never nested inside it, and offers to add it to a sibling `*.code-workspace` if one exists.
Tailored evening wrap — what shipped, what drafted, what stalled, goal-delta, and tomorrow's first move. Strict today-only scope; cheeky one-liner when the day is genuinely empty.
Cross-project work session. Triage active tasks, advance every project meaningfully, close what's done, log what mattered. Invoke when you have time to work across the whole portfolio rather than one focus area.
Set this month's themes and big-rocks goals — reads the full task board + last 2 weeks of sessions, then interviews you (2-3 rounds) before drafting themes. Run on the 1st of the month or whenever feels right.
基于 SOC 职业分类
| name | google-search-console |
| description | Pull real Google Search Console data (queries, impressions, clicks, CTR, average position, indexing status) for the configured site via `bun run dispatch google-search-console`. Use this instead of guessing what users search for, whenever you need SEO or search-traffic evidence. |
| disable-model-invocation | true |
Real search traffic data for the configured site. The GSC_SITE_URL env var names the Search Console property; the authenticated OAuth2 client lives under app/.config/ and is shared with other google- plugins* (currently google-page-speed) — one consent flow covers all.
Invoke as bun run dispatch google-search-console <cmd> [args...]. Output is always JSON on stdout — pipe through jq for filtering.
The command surface is a closed set: query, inspect, sites, auth. There is no overview, summary, totals, compare, or wow command — do not invent one. Rollups and period-over-period deltas are computed from query output with jq (see "Totals & WoW" under Common patterns).
query <start> <end> [--dim=query,page] [--page=<url>]Search-analytics query. Dates in YYYY-MM-DD. Default dimension is query (top search terms). Returns rows sorted by clicks.
--dim= picks the row shape; --page=<url> scopes the query to a single page (applies a dimension=page, operator=equals filter). The two are independent — you can run --dim=query --page=<url> to list the top queries for that one URL.
Flags accept both --flag=value and --flag value forms.
# Top 1000 queries last 7 days
mcp__plugin_agent-kevin_kevin__gsc_query 2026-04-13 2026-04-19
# Top URLs (not queries) last 30 days
mcp__plugin_agent-kevin_kevin__gsc_query 2026-03-20 2026-04-19 --dim=page
# Query × page pairs
mcp__plugin_agent-kevin_kevin__gsc_query 2026-04-13 2026-04-19 --dim=query,page
# Top queries for a single page (great for SERP diagnosis on a declining URL)
mcp__plugin_agent-kevin_kevin__gsc_query 2026-04-13 2026-04-19 --page=https://example.com/topic-keyword-a/
# Top 10 queries by clicks
mcp__plugin_agent-kevin_kevin__gsc_query 2026-04-13 2026-04-19 | jq '.[:10] | .[] | {keys:.keys[0], clicks, impressions, ctr, position}'
Row shape:
{ "keys": ["<query string>"], "clicks": 12, "impressions": 340, "ctr": 0.035, "position": 8.2 }
inspect <url>URL inspection — indexing status, last crawl time, mobile usability, rich-results. Use to diagnose why a specific page isn't ranking or getting traffic.
mcp__plugin_agent-kevin_kevin__gsc_inspect https://example.com/topic-keyword-e
sitesList verified properties accessible with the current credentials. Useful when query fails with a permission error — confirm the GSC_SITE_URL string matches exactly (including sc-domain: prefix for domain properties vs. full URL for URL-prefix properties).
mcp__plugin_agent-kevin_kevin__gsc_sites | jq '.[] | .siteUrl'
Totals & week-over-week (site-wide rollup for the current period vs prior period):
# Totals for a single window — sum across all pages
mcp__plugin_agent-kevin_kevin__gsc_query 2026-03-23 2026-04-19 --dim=page \
| jq '{clicks: (map(.clicks) | add),
impressions: (map(.impressions) | add),
ctr: ((map(.clicks) | add) / (map(.impressions) | add)),
avg_position: ((map(.position * .impressions) | add) / (map(.impressions) | add))}'
# WoW deltas — run the same jq against two adjacent windows, diff the results
mcp__plugin_agent-kevin_kevin__gsc_query 2026-02-23 2026-03-22 --dim=page \
| jq '{clicks: (map(.clicks) | add), impressions: (map(.impressions) | add)}'
Avg position must be impression-weighted (not a flat mean), or pages with few impressions will distort the number.
Find underperforming pages (high impressions, low CTR):
mcp__plugin_agent-kevin_kevin__gsc_query 2026-04-01 2026-04-19 --dim=page \
| jq 'map(select(.impressions > 100 and .ctr < 0.02)) | sort_by(-.impressions)'
SEO opportunity scan (pages ranking #11–20 with traffic):
mcp__plugin_agent-kevin_kevin__gsc_query 2026-03-20 2026-04-19 --dim=query,page \
| jq 'map(select(.position > 10 and .position < 20 and .impressions > 50)) | sort_by(-.impressions)'
Diagnose a specific page:
mcp__plugin_agent-kevin_kevin__gsc_inspect https://example.com/topic-keyword-c
queryinspectquery with --dim=page for traffic trend, then inspect for technical issuesIf commands fail with "Tokens not minted" — the one-time OAuth dance hasn't been run yet:
mcp__plugin_agent-kevin_kevin__google_auth
This opens a browser, captures Google's consent, and persists the refresh token to .kevin/secrets/google/google-tokens.json (alongside the OAuth client JSON, in the deny-gated secrets dir) — the same token is used by every google- plugin*, so you only run auth once, not once per plugin. You handle this once per install; Kevin should not normally need to run auth.