read-any-document (Canonizr)

Convert any document to markdown using Canonizr

Security Scan

The skill's instructions consistently describe using a local Canonizr CLI to convert documents to Markdown; there are no unexpected credential requests or install steps, though the SKILL.md assumes a canonical CLI/service is present without documenting installation.

Purpose & Capability

The name and description match the instructions (convert documents to Markdown via Canonizr). However, the skill expects a 'canonizr' CLI/service and optionally Docker to be available but the metadata lists no required binaries or install steps — this is a minor documentation mismatch rather than a functional incoherence.

Instruction Scope

Instructions are narrowly scoped to running the canonizr CLI (convert, health, up, down) and using --json. They do not instruct reading unrelated files or exfiltrating data. They do assume the user can start the pipeline (mentions Docker) but do not explain whether the pipeline is local or remote.

Install Mechanism

No install spec (instruction-only) — lowest risk. The skill does not attempt to download or install code itself.

Credentials

No environment variables, credentials, or config paths are requested; nothing disproportionate is asked for.

Persistence & Privilege

Skill is not always-enabled, does not request elevated persistence, and does not modify other skills or system-wide settings.

Assessment

This skill is coherent: it expects a local 'canonizr' pipeline/CLI to be present and tells the agent how to call it. Before installing or using it, verify where 'canonizr' comes from and whether it runs locally or contacts an external service (to avoid unintentionally uploading sensitive documents). If you plan to run the pipeline yourself, install canonizr and Docker from official/trusted sources and confirm the service's behavior (does it store or transmit files?). Also consider updating the skill metadata to declare the required binaries (canonizr and optionally docker) so the expectations are explicit.

Read any document using the Canonizr pipeline. Supports PDFs, images, office files, and more.

Using Canonizr

You can pass documents to Canonizr via CLI.

canonizr convert document.pdf

The output of the above command is a direct markdown transcript.

Failed attempts to convert a file return an error message. Depending on context you may wish to store the output in an adjacent .md file to avoid re-processing.

Supported file formats include PDFs, images (which will be transcribed), and office files.

Debugging

You can check if the Canonizr pipeline is running by using the CLI:

canonizr health

If the pipeline is not running then your user may not have started it. If you have docker permissions then you can start the service with:

canonizr up

And stop it with:

canonizr down

You can get a full JSON response with metadata (instead of just markdown) using the --json flag:

canonizr convert --json document.pdf