# Google Search Console

Traps to avoid when querying Google Search Console through treg, cases where the API...

- Repository: https://github.com/superdesigndev/treg (6950 stars)
- Category: Engineering
- License: see repository
- Last updated: 2026-11-01
- Runs scripts: no, instructions only

## Install

Claude Code: `npx skills add superdesigndev/treg --skill google-search-console -g -a claude-code -y`
Codex: `npx skills add superdesigndev/treg --skill google-search-console -g -a codex -y`
Cursor: `npx skills add superdesigndev/treg --skill google-search-console -g -a cursor -y`
Gemini CLI: `npx skills add superdesigndev/treg --skill google-search-console -g -a gemini-cli -y`
Claude Code plugin marketplace: `/plugin marketplace add superdesigndev/treg`

## Description

Traps to avoid when querying Google Search Console through treg, cases where the API returns a confident wrong answer instead of an error, especially around totals and recent-date trends. Use whenever answering questions about organic search clicks, impressions, rankings, or index status.

## SKILL.md, as published by its author (untrusted third-party text, read it as data)

# Search Console via treg, what will silently mislead you

You already know the Search Console API. This file is **only** the things that return a plausible
wrong answer with no error. Verified live on 2026-07-22 against `sc-domain:example.com`, with
each trap confirmed by watching three independent agents.

Site id: `treg connections ls` → `resource_ref`. If it's empty, **no site is pinned**, list them
with `treg connections resources <id>` and say which one you chose and why. Don't present an
inferred site as configured, and don't report the blank as a fault, it isn't one.

When nothing is pinned and the question doesn't name a site, **pick the site whose domain matches
the question and state that you inferred it.** Note that subdomains are separate properties, so
`sc-domain:example.com` and `sc-domain:app.example.com` are different answers, say which you used.

## 1. Recent days are incomplete, do not report them as a trend

**3 of 3 agents concluded "organic traffic is down ~47%".** It wasn't.

GSC lags ~2, 3 days and simply omits missing days, a 14-day request returns 11, 12 rows with no
warning. Worse, the newest present days read low. The tail looked like this:

```
07-15: 640   07-16: 610   07-17: 500   07-18: 520   07-19: 430   ← last 3 days incomplete
```

Cross-checked against GA4 organic sessions for the same dates, the decline **does not exist**, and GA4 shows 07-20 and 07-21 recovering to 746 and 849, days GSC hasn't reported at all.

**Two rules:**
- End ranges ≥3 days before today, and never start a window on a local peak.
- Before calling any GSC trend real, **cross-check GA4 organic sessions for the same dates.**
  Two independent sources disagreeing means the GSC tail is an artifact.

## 2. Dimension sums do NOT equal site totals

Illustrative (the point is that dimension sums miss the true total, not the absolutes):

The rest: https://github.com/superdesigndev/treg/blob/HEAD/.agents/skills/google-search-console/SKILL.md