monthly-report
Skill7 starsChanged 3 months agoMITView on GitHub
Produce a client-ready monthly search performance report with brand and non-brand split, winners, losers, and recommended actions. Use when asked for a monthly report, a recurring site summary, how the site performed over a period, or a client-facing performance update. Not for diagnosing a specific drop — that is traffic-decline-diagnosis.
8 other projects use a version of this.
Adopted by
Most recently taken up by
- tang1693/yunduan-ai-skillstheir own version, first seen Nov 8, 2026
- omarmfouad25/google-marketing-agenttheir own version, first seen Oct 25, 2026
- smorrel1/claude-code-skillstheir own version, first seen Oct 20, 2026
- apeabody007/Digitstheir own version, first seen Sep 4, 2026
- xyyandhtl/ClaudeSkillstheir own version, first seen Jun 17, 2026
- logly/mureotheir own version, first seen Apr 13, 2026
Other versions8 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 8
- logly/mureo★ 64same name
- claude-office-skills/claude-office-plugin★ 9same name
- omarmfouad25/google-marketing-agent★ 4same name
- smorrel1/claude-code-skills★ 2same name
- tang1693/yunduan-ai-skills★ 2same name
- apeabody007/Digits★ 1same name
- breethomas/bette-os★ 0same name
- xyyandhtl/ClaudeSkills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository7 other files it gives its agents.
- AGENTS.mdAGENTS.md
- .cursor/rules/gsc-mcp.mdcCursor rule
- ctr-gapsSkill
- gsc-recipesSkill
- index-healthSkill
- nonbrand-performanceSkill
- traffic-decline-diagnosisSkill
The file
We split this skill into the rules it gives an agent
9 rules. 0 are common, 3 are distinctive, and 6 are very specific to this project.
0
Common
Most of the 6 edited versions we read keep these.
3
Distinctive
This author's own choices. Worth a look.
6
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 12
Call list_sites to resolve the exact site_url.
- Distinctive: line 14
Default window: the last 28 days ending 3 days ago, compared against the 28 days immediately before.
- Distinctive: line 38
Ask the user for the brand terms if this is the first report of the conversation.
- Very specific: line 68
Do not take winners and losers from one clicks_delta desc call.
- Very specific: line 70
Call compare_periods twice per dimension:
- Very specific: line 96
After every call, read truncated, scan_capped, and ordering:
- Distinctive: line 139
Do not attribute causes the data cannot support.
- Very specific: line 140
Always read truncated, scan_capped, and ordering on every compare_periods / query_search_analytics response.
- Very specific: line 141
Never derive brand/non-brand share or change from a capped grouped query list.
“Common” is measured: kept by at least 6 in 10 of the 6 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Monthly Search Report
A report someone can forward to a client without editing. Numbers, what changed, and what to do about it.
Step 1 — Property and window
Call list_sites to resolve the exact site_url.
Default window: the last 28 days ending 3 days ago, compared against the 28 days immediately before. Twenty-eight days rather than a calendar month keeps weekday counts equal — a calendar month comparison mixes a 4-weekend month with a 5-weekend one and produces fake swings.
State the exact dates in the report.
Step 2 — Overall movement
compare_periods
site_url: <resolved>
period_a_start / period_a_end: <previous 28 days>
period_b_start / period_b_end: <recent 28 days>
dimensions: [] # totals onlyGives clicks, impressions, CTR and position for both periods plus deltas. Note the field semantics:
ctr_delta_ppis in percentage points (2.5 means 2.5pp, not 0.025)position_improvedis a boolean — a smaller position number is better, and this field has already worked that outonly_inmarks rows present in one period only; those rows omit position deltas because a comparison would be meaningless
Step 3 — Brand vs non-brand split
KPI totals (clicks, share, change) come from aggregate calls: omit dimensions. Do not sum a dimensions: ["query"] top-N for these numbers — row_limit would silently drop traffic and truncated cannot recover the missing total.
Ask the user for the brand terms if this is the first report of the conversation.
Non-brand totals:
compare_periods
...same dates...
dimensions: []
dimension_filter_groups: [{
"groupType": "and",
"filters": [{"dimension":"query","operator":"excludingRegex","expression":"<brand pattern>"}]
}]Branded totals:
compare_periods
...same dates...
dimensions: []
dimension_filter_groups: [{
"groupType": "and",
"filters": [{"dimension":"query","operator":"includingRegex","expression":"<brand pattern>"}]
}]This split is what separates a real report from a vanity one. Branded traffic tracks brand awareness; non-brand tracks search performance. Query-level detail stays in Step 4.
Step 4 — Winners and losers
Do not take winners and losers from one clicks_delta desc call. That ranking only returns the biggest gains; the biggest drops sit at the other end of the list and are cut by row_limit.
Call compare_periods twice per dimension:
Winners (query):
compare_periods
...same dates...
dimensions: ["query"]
sort_by: clicks_delta
sort_order: desc
row_limit: 5Losers (query):
compare_periods
...same dates...
dimensions: ["query"]
sort_by: clicks_delta
sort_order: asc
row_limit: 5Repeat both calls with dimensions: ["page"]. Pages that lost clicks are usually more actionable than queries, because a page is something you can go and fix.
After every call, read truncated, scan_capped, and ordering:
truncated=truemeans more rows exist beyondrow_limit.scan_capped=truemeans even the 25,000-row scan was incomplete, so the top-N may be missing candidates.- Confirm
orderingmatches what you asked for before writing the table.
Also collect:
- Queries/pages with
only_in: "b"— newly appearing - Queries/pages with
only_in: "a"— disappeared
Step 5 — Write the report
# Search Performance — <site>
<period B dates> vs <period A dates>
## Summary
One paragraph in plain language. Lead with non-brand, since that is the SEO result.
## Numbers
| Metric | Previous | Current | Change |
| Clicks / Impressions / CTR / Avg position |
## Brand vs non-brand
Table with clicks, share, and change for each.
## What grew
Up to 5 rows, each with a one-line reason if the data suggests one
(position improved, new query appeared, CTR rose at the same position).
## What declined
Up to 5 rows. For each, say whether position fell, impressions fell,
or CTR fell — the three have different causes and different fixes.
## Recommended actions
Three at most, ordered by expected impact. Each names a specific page or query.Accuracy rules
- Give exact date ranges. "Last month" is ambiguous and clients will compare against their own dashboard.
- Round rates sensibly: CTR to one decimal as a percentage, position to one decimal.
- Do not attribute causes the data cannot support. A ranking drop is visible; why it dropped is not in Search Console.
- Always read
truncated,scan_capped, andorderingon everycompare_periods/query_search_analyticsresponse. Do not treat a top-N as complete when either flag is true. - Never derive brand/non-brand share or change from a capped grouped query list. Use the aggregate (no-dimension) calls from Step 3. Anonymized
(other)queries can still make the two halves miss the site total — say so. - API rows are top rows only, so category totals will not sum to the site total. Say so if you show both.
- Impressions moving without clicks moving usually means a position or SERP-feature change, not a content problem — check position before recommending a rewrite.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

