hooks — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited hooks (Hook) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
A skills library for product data scientists working with Claude.
Reusable, opinionated skills you can invoke in Claude Code, Cursor, and other AI coding harnesses for the analytics work you actually do every day — experiments, cohorting, SQL modeling, regression, causal inference, and stakeholder readouts.
Built in the spirit of affaan-m/ECC: skills are the primary surface, descriptions trigger discovery, and every skill ships with reference docs, examples, and reusable scripts.
Most "analytics templates" are either generic markdown checklists or one-off notebooks that rot fast. This library treats DS workflows as first-class skills:
SKILL.mdanalytics-skills/
├── skills/ # Primary surface — 15 day-to-day DS skills
├── agents/ # Specialized reviewers (experiment, SQL, readout)
├── commands/ # Slash commands for fast invocation
├── rules/ # Always-follow style guides
├── hooks/ # Pre/post tool automations
├── templates/ # Notebook & readout templates
├── manifest.json # Catalog of all skills + metadata
├── INSTALL.md # Setup for Claude Code, Cursor, manual
└── CONTRIBUTING.md # Guide for adding new skills| Skill | What it does | When Claude uses it |
|---|---|---|
ab-test-design | Power, MDE, sample size, design checklist | "Design an A/B test for…" |
ab-test-analysis | Significance testing, CIs, readout | "Analyze this experiment…" |
cohort-analysis | Retention curves, cohort tables | "Show retention by signup cohort…" |
funnel-analysis | Step conversion, drop-off diagnosis | "Build a funnel for signup → activation…" |
metric-definition | Metric specs, guardrails, owner | "Define a North Star for…" |
modular-sql-ctes | Staging/intermediate/fact CTE patterns | "Refactor this SQL…" |
data-quality-audit | Null/dup/freshness/schema checks | "Audit this table for data quality…" |
logistic-regression | Binary outcome modeling, calibration | "Model churn probability…" |
linear-regression | Continuous outcome, coefficient reading | "Predict revenue from…" |
survival-analysis | Time-to-event with censoring | "Model time-to-churn…" |
stakeholder-readout | Insight write-up structure | "Write a readout for…" |
warehouse-query-optimization | Snowflake-specific perf tuning | "This query is slow on Snowflake…" |
sql-query-review | Static anti-pattern check + optimized rewrite | "Review/optimize this query…" |
sql-correctness-review | Logic audit: dupes, fanout, joins, NULLs, CASE | "These numbers look wrong…" |
causal-inference | DiD, matching, IV decision framework | "Estimate causal impact of…" |
git clone https://github.com/<your-org>/analytics-skills.git
cd analytics-skills
# Install for Claude Code
bash install.sh --target claude
# Install for Cursor (project-local)
bash install.sh --target cursor --project-dir /path/to/your/projectSee INSTALL.md for all install paths, including manual copy and partial installs.
Once installed, Claude will auto-trigger skills based on intent. You can also invoke directly:
"Use the ab-test-analysis skill to read out the experiment in /data/exp_123.csv"
"Apply modular-sql-ctes to refactor this query: <paste>"
"Run a data-quality-audit on fct_orders"Each skill prompts Claude through a consistent workflow:
Skills are written against this stack. Where logic is stack-specific, alternatives are noted in each skill's reference.md.
MIT. See LICENSE.
See CONTRIBUTING.md. New skills welcome — follow the template and quality checklist.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.