jeremylongshore/tons-of-skills-marketplace

web-analytics

Analyze Umami or GA4 traffic with evidence-backed specialist reviews and deliver a concise portfolio pulse, decision brief, or deep dive.

View source

How to use

  1. Copy the installation command and review the source package before installing it in your own agent environment.
  2. Check the source document for required tools, credentials and supported hosts.
  3. Describe your input and desired deliverable; review the output against the source material.
npx skills add https://github.com/jeremylongshore/tons-of-skills-marketplace --skill web-analytics

Before you start

  • Installation does not include a model subscription, API allowance or permission to access your files.
  • Repository stars and installation counts describe activity, not an independent quality or security test.
Original skill document

Rendered from the source repository. Headings, examples, code, tables, links, and referenced images are preserved.

Web Analytics Intelligence

Overview

Turn portfolio analytics into a decision-ready report by collecting a bounded dataset, preserving the comparison window, separating observed facts from hypotheses, and routing deeper requests through specialist agents. This is a push-based analysis workflow, not a dashboard replacement.

Read the operating contract before making authenticated requests or delivering a report.

Prerequisites

  • Configure sites and thresholds in ${CLAUDE_PLUGIN_ROOT}/references/site-registry.md.
  • Store Umami credentials in the operator's approved secret store; this installation uses

UMAMI_PASSWORD from ~/.env.

  • Confirm /email or /slack independently before requesting those delivery channels.
  • Use ${CLAUDE_PLUGIN_ROOT}/references/reporting-tiers.md and

${CLAUDE_PLUGIN_ROOT}/references/interpretation-guide.md for medium and full reports.

Authentication and Safety

  1. Confirm the analytics base URL is the expected operator-controlled host before sending a

credential.

  1. Load UMAMI_PASSWORD only for the command that needs it. Never print, log, persist, or pass

the password or bearer token to a specialist agent.

  1. Use curl --fail-with-body --silent --show-error; stop if authentication fails or the token

is empty.

  1. Treat email and Slack delivery as external side effects. Preview the destination and report,

and obtain confirmation when the user has not explicitly requested delivery.

  1. Never send messages, modify analytics configuration, or write baseline state during a read-only

request. A full-tier memory update requires an explicit writable scope.

Workflow

1. Parse the Request

ParameterDefaultAccepted values
Tierminimini, medium, full
SiteallRegistry name or all
Period7dtoday, yesterday, 7d, 30d, mtd, qtd
Deliveryconsoleconsole, email, slack, all
Compareprior equivalentAn explicit comparison window

Use defaults when they preserve the user's intent. Ask before continuing if the site, time window, or external destination would materially change the result.

2. Load Configuration

Read only the files needed for the selected tier:

  1. ${CLAUDE_PLUGIN_ROOT}/references/site-registry.md for site IDs, baselines, and thresholds.
  2. ${CLAUDE_PLUGIN_ROOT}/references/mcp-tool-reference.md for supported data operations.
  3. ${CLAUDE_PLUGIN_ROOT}/references/reporting-tiers.md for output contracts.
  4. ${CLAUDE_PLUGIN_ROOT}/references/interpretation-guide.md for evidence language.

3. Collect the Minimum Dataset

For direct Umami access, authenticate and fail closed:

bash
set -euo pipefail
source ~/.env
analytics_url="https://analytics.intentsolutions.io"
token=$(curl --fail-with-body --silent --show-error "${analytics_url}/api/auth/login" \
  -X POST -H "Content-Type: application/json" \
  -d '{"username":"admin","password":"'"${UMAMI_PASSWORD}"'"}' | \
  python3 -c "import json,sys; print(json.load(sys.stdin).get('token',''))")
test -n "${token}" || { printf 'Umami authentication returned no token\n' >&2; exit 1; }
curl --fail-with-body --silent --show-error \
  "${analytics_url}/api/websites/<site-id>/stats?startAt=<start-ms>&endAt=<end-ms>&compare=prev" \
  -H "Authorization: Bearer ${token}"
unset token UMAMI_PASSWORD

Record the source, site, timezone, exact start and end timestamps, comparison window, and any missing endpoint. Do not infer missing values as zero.

4. Route by Tier

  • Mini: Collect aggregate stats and active visitors inline. Return totals, per-site metrics,

comparison deltas, and one material signal in at most 15 lines.

  • Medium: Use Agent to run data-collector; after it returns, run traffic-intelligence,

content-seo, and anomaly-detector concurrently. Give reporting-narrative their outputs, the exact period, and the requested delivery format.

  • Full: Run data-collector for aggregate, event, technology, and geography data. Then run all

five analysis specialists concurrently, pass their claims to verification-agent, and compile only verified or explicitly qualified findings with reporting-narrative.

Agent definitions live under ${CLAUDE_PLUGIN_ROOT}/agents/. Give agents collected results, not credentials. Bound every assignment to the requested sites and period. If an agent fails, report that coverage gap and continue only when the remaining evidence can support the requested output.

5. Verify and Deliver

Before delivery, apply the operating contract:

  1. Recalculate headline deltas from the raw totals.
  2. Label each statement as observation, comparison, hypothesis, or recommendation.
  3. Check anomaly claims against baselines and low-volume noise.
  4. Name unavailable sources and incomplete windows.
  5. Preview external destinations, then invoke /email or /slack only when authorized.

For a full-tier report, run memory-agent only if the user authorized baseline-state updates. Persist the period, source, and evidence receipt so a later report can reproduce the comparison.

Error Handling

Stop without exposing secret material when authentication fails. For partial endpoint, site, agent, or delivery failures, retain successful evidence, name the exact coverage gap, and avoid complete- portfolio or trend claims that the remaining data cannot support. The operating contract defines the required fallback for each failure class.

Output

Return the tier, sites, exact period, comparison window, sources consulted, coverage gaps, and delivery result. Lead with the strongest verified signal, show the supporting metrics, distinguish hypotheses from facts, and end with prioritized actions that each have an owner or next check.

Examples

  • /analytics --period=today produces a mini console pulse for every configured site.
  • /analytics medium --site=tonsofskills --period=7d explains material traffic and content shifts.
  • /analytics full --period=30d --email previews an evidence-checked deep dive before email delivery.

Resources

  • Operating contract — auth, evidence, failure, and delivery gates.
  • ${CLAUDE_PLUGIN_ROOT}/references/site-registry.md — configured properties and thresholds.
  • ${CLAUDE_PLUGIN_ROOT}/references/reporting-tiers.md — detailed report schemas.
  • ${CLAUDE_PLUGIN_ROOT}/references/interpretation-guide.md — analytical voice and caveats.
  • ${CLAUDE_PLUGIN_ROOT}/references/delivery-channels.md — email and Slack adapters.
from this repository

More skills

All skills
jeremylongshore
Community

agent-context-loader

'Execute proactive auto-loading: automatically detects and loads agents.md files. Use when appropriate context detected. Trigger with relevant phrases based on skill purpose. '

1h installs
2
GitHub stars
2.9K
Updated
Nov 8
jeremylongshore
Community

agent-creator

'Create production-grade agent .md files aligned with the current Anthropic subagent contract. Also validates existing agents against the marketplace compliance rules. Use when building custom subagents, reviewing agent quality, or creating parallel agent architectures for orchestrator skills. Trigger with "/agent-creator", "create an agent", "build a subagent", or "validate my agent". Make sure to use this skill whenever creating agents/.md files for plugins or standalone use. '

1h installs
2
GitHub stars
2.9K
Updated
Nov 8
jeremylongshore
Community

agent-patterns

'Execute this skill should be used when the user asks about "SPAWN REQUEST format", "agent reports", "agent coordination", "parallel agents", "report format", "agent communication", or needs to understand how agents coordinate within the sprint system. Use when appropriate context detected. Trigger with relevant phrases based on skill purpose. '

1h installs
2
GitHub stars
2.9K
Updated
Nov 8
jeremylongshore
Community

agent-safety-preflight

Generate a local repository risk receipt before Claude Code or other AI-agent edits. Use when the user asks to prepare a repository for agent changes, check for risky hooks or credential-writing automation, or run the /agent-preflight command. Trigger with "run agent preflight", "check this repo before agent edits", or "/agent-preflight".

1h installs
2
GitHub stars
2.9K
Updated
Nov 8