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 clamp-sh/analytics-skills --skill metric-context-and-benchmarksgit clone --depth 1 https://github.com/clamp-sh/analytics-skillsWrote 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/clamp-sh/analytics-skills/metric-context-and-benchmarks)<a href="https://agentmods.dev/skills/clamp-sh/analytics-skills/metric-context-and-benchmarks"><img src="https://agentmods.dev/badge/skills/clamp-sh/analytics-skills/metric-context-and-benchmarks/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/clamp-sh/analytics-skills/metric-context-and-benchmarks"><img src="https://agentmods.dev/badge/skills/clamp-sh/analytics-skills/metric-context-and-benchmarks.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 | $0.00092 | $0.04139 |
| Opus 5 | $0.00046 | $0.02070 |
| Sonnet 5 | $0.00018 | $0.00828 |
| Haiku 4.5 | $0.00009 | $0.00414 |
Grade A, and why
metric-context-and-benchmarks 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 11d ago.
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 — 315 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Metric context and benchmarks
The reference skill for answering "is this number good, bad, or noise?" correctly. Almost every wrong answer to that question comes from quoting a cross-industry average when the user's industry-specific number is 3× different, or quoting a benchmark without checking if the sample size even supports the comparison.
Read analytics-diagnostic-method first. Sample-size discipline applies here more than anywhere: benchmarks are meaningless if your data is noise.
How to use this skill
- Load
analytics-profile.md. The profile's model + industry row picks the right benchmark. - Check the sample size. If the user's number is computed on too few observations, the comparison to any benchmark is moot. Say so and stop.
- Look up the relevant benchmark with its qualifiers (year, source, population, definition).
- Answer in the form: "Your X is Y. Benchmark for [specific population] is Z [from source, year, n=]. Your number is [above / at / below] benchmark by W percentage points."
- Flag metric-specific traps (e.g. GA4 bounce rate is not UA bounce rate; last-click CVR is not the full story).
The honesty rule
Never quote a benchmark as if it's a target. Benchmarks describe populations; they don't tell you what your number should be. A business can have a below-benchmark CVR and a healthy LTV:CAC if their AOV or retention is elevated. A business can have a 3× benchmark CVR and still be unprofitable. Always connect the benchmark back to the user's actual unit economics.
Metric-by-metric
Bounce rate / engagement rate
GA4 definition (authoritative, from support.google.com/analytics/answer/12195621):
- Engaged session = session lasting longer than 10 seconds, OR with a conversion (key) event, OR with 2+ page/screen views.
- Engagement rate = engaged sessions / total sessions.
- GA4 bounce rate = 100% − engagement rate.
Critical: GA4 bounce rate is NOT the same as Universal Analytics bounce rate. UA bounce rate was single-pageview sessions. In GA4, a single-pageview session with >10s dwell is engaged (not bounced). So pre-2023 blog posts quoting "50% bounce rate is average" are measuring a different thing.
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.
- 11d ago First seen · 315 lines · 92 tokens per session scan A 74626d7f7c57
metric-context-and-benchmarks is a skill published in the GitHub repository clamp-sh/analytics-skills (8 stars, last pushed 3mo ago), licensed MIT. It adds 92 tokens to every session and 4,139 once invoked, about $0.0005 per session on Opus 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 2026-08-31.
Other skills, from other repositories
treat
Prune bloated session with a prescription. Removes progress ticks, stale reads, duplicate content, and more.
guard
Protect Claude Code sessions from context overflow by running a background daemon that monitors session size and auto-prunes before compaction hits. Use when the user says "guard", "protect session", "context getting long", "prevent compaction", "session management", or is running agent teams that need continuous…
cozempic-doctor
Run health checks on Claude Code configuration and sessions. Use when troubleshooting Claude Code issues. (cozempic — does not shadow Claude Code's built-in /doctor).
diagnose
Analyze Claude Code session bloat — shows token count, context usage %, and bloat breakdown. Use when the user asks about session size, context usage, or when you notice the context window is getting full.
reload
Treat the current session and auto-resume in a new terminal window.
radin-execute
Work through a project's whole backlog: prioritize every task, execute each via a sub-agent, commit after each. Use when the user wants the entire backlog processed ("work through my backlog"), not one named task. Delegates all implementation to sub-agents; clarifies ambiguity by asking the user rather than guessing.