AI skills, ranked by what actually works in production.

Use when extracting structured fields from invoices or documents into a fixed schema — applies the org's canonical mappings (vendor name to canonical id, description to category code, money to cents, status enum) the base can't infer, so the output matches downstream systems instead of echoing raw strings. Template: replace with your schema and mappings.

Answer strictly from the provided sources, cite every claim, and refuse when the answer isn't there — instead of guessing from training knowledge.

Use when extracting named entities and normalizing each to the org's canonical id + type code (a registry the base can't know) rather than raw surface mentions — so the same real-world entity resolves to one id downstream. Template: replace with your entity registry and codes.

Use when harmonizing clinical results into a system's canonical units — applies the org's target units (mmol/L, umol/L, g/L, Celsius, kg), per-analyte rounding conventions, and ITS reference ranges for H/L/N flagging, instead of whatever unit the source used. Template: replace with your units and ranges.

Use when writing commit messages in a repo with a custom commit format — applies the org's convention ([TICKET] AREA: summary with arbitrary AREA codes) the base won't guess, so commits are consistent and parseable. Template: replace with your format.

Use when extracting contract clauses into legal-ops systems — applies the org's clause-type taxonomy, governing-law code set, term normalization (month-to-month, initial terms), and liability-cap classification the base can't guess. Template: replace with your taxonomy.

Use when assigning an SLA priority (P1-P4) to a support ticket — applies the org's priority matrix (severity x plan tier) and overrides (security always P1, VIP raises one level) the base can't infer from the ticket alone. Template: replace with your matrix and overrides.

Use when screening candidates against a rubric — applies the org's bars (minimum experience, required skills, work authorization, the advance/referral bar, role-on-hold) to assign a stage consistently. Template: replace with your rubric.

Use when an agent builds tool-call arguments from a user request — encodes the org's call policy beyond the bare schema (money in cents, required idempotency keys, priority/status enums, E.164 phone formatting, caps, and abstain-don't-fabricate on a missing required arg). Template: replace with your tool contracts.

Use when approving or denying a refund or return — encodes the org's refund windows and overrides (e.g. 30 days for goods, 14 for digital, defective overrides the window, large refunds escalate, pre-delivery is a cancellation) that override generic reasonableness. Template: replace with your refund policy.

Orchestrate a multi-agent dev/agency team with shared memory, handoffs, and OpenClaw sessions. Use when (1) First-time onboarding — follow references/OPENCLAW_TEAM_SETUP_GUIDE.md (guided variables + master prompt) and references/SKILL-SETUP.md, (2) Running specialized agents (PM, dev, QA, security) on customer or Shopware tasks, (3) Setting up shared team folders and SOUL.md identities, (4) sessions_spawn/sessions_send or control channels (e.g. Telegram), (5) Solo-founder multi-agent pattern fro

support ticket triage decisions per internal policy

Use when deciding whether and where to escalate a support ticket — encodes the org's escalation thresholds and targets (Sev-1 to oncall, refunds over a limit to a manager, Enterprise churn to retention, legal/press to legal) the model can't guess from the ticket alone. Template: replace with your thresholds and targets.

SKU canonicalization decisions per internal policy

SKU canonicalization decisions per internal policy

support ticket triage decisions per internal policy

SKU canonicalization decisions per internal policy

invoice field extraction decisions per internal policy

invoice field extraction decisions per internal policy

SKU canonicalization decisions per internal policy