redacta
by nickjlamb
Coding Agents & IDEsBrief
Redacta pseudonymises medical documents — replacing patient identifiers (NHS numbers, dates of birth, postcodes.
Description
What This Skill Does
Redacta pseudonymises medical documents by detecting and replacing patient identifiers with labelled tokens such as [NHS_NUMBER], [PATIENT_NAME], and [DATE_OF_BIRTH]. It handles both structured identifiers via regex patterns and contextual ones via clinical reasoning. Clinical content and clinician names are preserved by default.
Token labels preserve document structure and clinical meaning, so downstream AI can still interpret the content without ever seeing real patient data.
When to Use It
- Preparing patient letters for AI summarisation
- Anonymising GP referral letters before third-party processing
- Stripping identifiers from discharge summaries for clinical audit
- Cleaning medical notes prior to research data review
- Redacting correspondence before sharing with external AI tools
Example Workflow
User asks: Please redact this letter: 'Dear Mrs Patricia Hartley, DOB: 14/03/1952, NHS Number: 943 476 5919, Tel: 0113 278 4532...'
- 1
Scan text for structured identifiers using regex patterns (NHS numbers, dates, postcodes, phone numbers)
- 2
Detect contextual identifiers such as patient name, address, and age using clinical document reasoning
- 3
Replace each identifier with its labelled token consistently throughout the document
- 4
Preserve clinical content, appointment dates, and clinician names
- 5
Return the pseudonymised document followed by a redaction report listing each token type, count, and line
Pseudonymised letter with all patient identifiers replaced by tokens, plus a redaction report summarising 6 items replaced and confirming clinical content preserved
Installation
clawhub install nickjlamb/redactaRun the CLI command, or paste the prompt into Claude Code, Cursor, Codex or whichever coding agent you use.