Give Claude Code full context on the AgileTrust Tokenization API — endpoints, auth, encoding modes, algorithm trade-offs, tweak rules, and PII-specific recipes — so your team can integrate in minutes instead of hours.
Installation
The skill is installed once per machine. After installation, Claude Code automatically loads it whenever you ask about tokenization, FPE, or AgileTrust integration.
Requires curl and tar (pre-installed on macOS and most Linux distros):
curl -L https://tokenization.qa.agiletrust.io/agiletrust-tokenization-skill.tar.gz \
| tar -xz -C ~/.claude/skills/
Replace https://tokenization.qa.agiletrust.io with your production docs URL when deploying.
mkdir -p ~/.claude/skillstar -xzf agiletrust-tokenization-skill.tar.gz -C ~/.claude/skills/ls ~/.claude/skills/agiletrust-tokenization/SKILL.mdTo install the skill for a specific project (all team members on the repo benefit when they clone it):
mkdir -p .claude/skills
curl -L https://tokenization.qa.agiletrust.io/agiletrust-tokenization-skill.tar.gz \
| tar -xz -C .claude/skills/
# Commit .claude/skills/agiletrust-tokenization/ to your repo
Usage
After installation, just ask Claude Code naturally. The skill triggers automatically when you mention tokenization, FPE, AgileTrust, or PII data types. You can also invoke it directly.
/agiletrust-tokenization
This loads the full skill context and prints a usage summary.
POST /tokenize with encoding: "numeric". The hyphen and check digit will be preserved automatically — you just pass the full RUT string. Here's a working snippet: tokenize("13301430-6", encoding="numeric") → "35240589-5". Note: recalculate the check digit after detokenizing if your system validates Modulo-11.@) and preserve the domain. For example, split user@example.com, tokenize user with encoding: "utf8" and tweak "656d61696c0000", then reassemble as token@example.com. The @ and .com stay intact for routing.The skill activates automatically when you use phrases like:
Contents
Main skill file: setup, quick start, encoding/algorithm selection, tweak rules, error handling, code samples, and use-case table. Always loaded when skill triggers.
Full endpoint specs, request/response schemas, all field types and constraints, format-preservation examples, and algorithm-specific limits.
FF3-1 vs FF1 detailed comparison: pros/cons, capability matrices by encoding, when-to-choose guide, and links to NIST standards.
Per-PII recipes: email, RUT/DNI, credit card (PAN), phone, full name, date, IBAN, IP address. Each includes curl + Python examples and caveats.
All HTTP status codes (400–503) with body examples, reproduction curl commands, and fixes.
Runnable scripts: tokenize.sh (bash/curl), tokenize.py (Python requests), tokenize.js (Node.js fetch). Copy and run directly.
Bundled OpenAPI 3.0.3 spec for offline reference. Identical to the live API Reference.