Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add BoxLab-Ltd/yandex-metrica-mcp --skill traffic-dropgit clone --depth 1 https://github.com/BoxLab-Ltd/yandex-metrica-mcpWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/boxlab-ltd/yandex-metrica-mcp/traffic-drop)<a href="https://agentmods.dev/skills/boxlab-ltd/yandex-metrica-mcp/traffic-drop"><img src="https://agentmods.dev/badge/skills/boxlab-ltd/yandex-metrica-mcp/traffic-drop/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/boxlab-ltd/yandex-metrica-mcp/traffic-drop"><img src="https://agentmods.dev/badge/skills/boxlab-ltd/yandex-metrica-mcp/traffic-drop.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00091 | $0.01376 |
| Opus 5.5 | $0.00036 | $0.00550 |
| Sonnet 5.5 | $0.00018 | $0.00275 |
| Haiku 4.5 | $0.00009 | $0.00138 |
Grade A, and why
traffic-drop scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 109 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Traffic drop investigation
A drop is a claim to test, not a fact to explain. Most "drops" are robots, calendar effects, a broken counter or one segment — find which before naming a cause. Work through the steps in order and stop as soon as the evidence answers the question.
0. Know your sources
- Metrica tools (
run_comparison,run_timeseries,run_report,describe_counter) are required. Without them, say so and stop. - Webmaster tools (
get_hosts,get_indexing,search_queries, …) are optional. If they are not available, do the Metrica steps, then list the search hypotheses you could not check and how the user can check them (install the yandex-webmaster-mcp plugin, or look in Webmaster: Indexing → Pages in search, Diagnostics). - If a
metrica-hygieneorwebmaster-auditskill is available, its rules apply here too.
1. Confirm the drop
run_comparisonof the period in question (B) against a baseline (A) of the same length and weekday mix, both complete (end yesterday or earlier),filtersA/filtersB=ym:s:isRobot=='No', metricsym:s:visits,ym:s:users.- Also compare the same weeks a year back (shift by 364 days) when the baseline might be seasonal (holidays, summer, New Year).
- Give absolute numbers. If the change is within noise (a few percent, or a small base), say there is no real drop and stop. If the unfiltered drop is large but people are flat, it was robots — say so and stop.
2. Date it
run_timeseries daily over ~6 weeks (robots excluded). A step on one day
points to a release, a tracking change or a search event; a slope points
to rankings, demand or seasonality. Note the first day of the change.
3. Find where it is
Break the comparison down one dimension at a time and compute each row's share of the total delta (row delta / total delta):
ym:s:lastsignTrafficSource— which source moved;- inside that source:
ym:s:lastsignSearchEngineRootfor search,ym:s:lastsignUTMSource/ym:s:lastsignUTMCampaignfor ads,ym:s:refererfor links; ym:s:startURLPath(landing page) — which pages lost entries;ym:s:deviceCategory,ym:s:regionCountryNameif still unclear.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 109 lines · 91 tokens per session scan A 40fb63ef603b
traffic-drop is a skill published in the GitHub repository BoxLab-Ltd/yandex-metrica-mcp (15 stars, last pushed yesterday), licensed MIT. It adds 91 tokens to every session and 1,376 once invoked, about $0.0004 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-11-06.
traffic-drop compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| traffic-dropBoxLab-Ltd/yandex-metrica-mcp | Skill | 15 | yesterday | MIT | A | 91 | Claude Code |
| 21st-ui21st-dev/magic-mcp | Skill | 6.4k | 1mo ago | ISC | A | 63 | Claude Code |
| testrail-qauarlouski/testrail-mcp-server | Skill | 44 | 4d ago | Apache-2.0 | A | 70 | Claude Code, Codex |
| usage-reportergetpapi/papi | Skill | 11 | yesterday | MIT | A | 57 | Claude Code |
| slack-send-safelyjtalk22/slack-mcp-server | Skill | 32 | 3d ago | MIT | A | 65 | Claude Code, Codex |
| papi-buildgetpapi/papi | Skill | 11 | yesterday | MIT | A | 26 | Claude Code |
| deployment-completeness-auditgetpapi/papi | Skill | 11 | yesterday | MIT | A | 44 | Claude Code |