Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add dambuchs/redact-pdf-mcp --skill anonymize-before-aigit clone --depth 1 https://github.com/dambuchs/redact-pdf-mcpWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/dambuchs/redact-pdf-mcp/anonymize-before-ai)<a href="https://agentmods.dev/skills/dambuchs/redact-pdf-mcp/anonymize-before-ai"><img src="https://agentmods.dev/badge/skills/dambuchs/redact-pdf-mcp/anonymize-before-ai/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/dambuchs/redact-pdf-mcp/anonymize-before-ai"><img src="https://agentmods.dev/badge/skills/dambuchs/redact-pdf-mcp/anonymize-before-ai.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00089 | $0.00811 |
| Opus 5.5 | $0.00036 | $0.00324 |
| Sonnet 5.5 | $0.00018 | $0.00162 |
| Haiku 4.5 | $0.00009 | $0.00081 |
Grade A, and why
anonymize-before-ai scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured today.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 51 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Anonymize, work, then put the names back
The point of this workflow is that the real names never enter the conversation. One careless read of the original file undoes it, so follow these steps in order.
1. Do not open the original
Do not Read, cat, preview, OCR or summarize the user's file before step 2. Pass its path to the tool and nothing else. If the user pasted the document's text into the chat instead of giving a file, tell them the names are already in the conversation, and suggest they save it as a PDF and give you the path next time.
2. Pseudonymize it
Call redact_pdf_and_wait with:
file_path: the user's PDF, JPEG or PNG. Pass images as they are; don't convert them to PDF.mode:"pseudonymize".
Every person, organization, address, email, phone number, date, IBAN and card number becomes a consistent tag such as [Person 1] or [Org 2], in the document's own language. The same person keeps the same tag on every page. National IDs and custom terms are blacked out instead.
The result gives you:
document_text: the pseudonymized document as text. Long documents returndocument_text_pathinstead; reading that file is fine, because it only contains tags.output_path: the pseudonymized PDF.correspondence_path: the tag → original table. Never open, read or quote it. It is the one file that re-identifies the document.
If there is no API key, call try_demo with the same file_path. It does the first page for free. Then tell the user a free account gets its first document (up to 5 pages) done in full. If the user then gives you a key in the conversation, call set_api_key with it once and carry on; don't repeat the key back.
3. Do the work on the tags
Do what the user asked using document_text only. Keep every tag exactly as written, brackets included, so it can be swapped back. Don't guess who a tag refers to, and don't invent names.
Write your result to a .md, .txt or .docx file in the user's folder.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- today First seen · 51 lines · 89 tokens per session scan A da9f3218505d
anonymize-before-ai is a skill published in the GitHub repository dambuchs/redact-pdf-mcp (1 stars, last pushed today), licensed MIT. It adds 89 tokens to every session and 811 once invoked, about $0.0004 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-19.
Other skills, from other repositories
pdf-processing
Extract text from PDFs, fill forms, and merge documents.
kernelcad-drawings
Export 2D engineering-drawing sheets (SVG, or a printable PDF on a standard sheet) from any model or assembly — third-angle front/top/left + isometric views with hidden-line removal, overall dimensions, and a title block. Use when the deliverable is a human-readable fabrication or review drawing.
agoragentic-anydoc-document-evidence
Convert a local Word, PowerPoint, spreadsheet, OpenDocument, RTF, EPUB, CSV, or text-based PDF into bounded Markdown and an Agoragentic evidence handoff. Use when an agent needs document contents plus runtime-verified provenance, explicit evidence coverage, known-loss warnings, and a pending-or-incomplete parse…
contract-analyzer
Extract contract parties, clauses, dates, and model-assessed risks using processdocument. Use when the user supplies contract text, Markdown, or a PDF/image for analysis.
pdf-invoice-extractor
Extract typed invoice data from supplied text, Markdown, or a PDF/image through processdocument. Use for invoice totals, line items, due dates, and extraction anomalies.
book-mirror
Take any book (EPUB/PDF), produce a personalized chapter-by-chapter analysis with two-column tables. Left column preserves the chapter content; right column maps every idea to the reader's actual life using brain context. Output is a single brain page at media/books/ -personalized plus an optional PDF via brain-pdf.