Matter CLI

Manage and interact with your Matter reading library: browse, search, save, tag, archive, highlight, summarize, and track reading progress via Matter CLI.

Security Scan

The skill is internally consistent with its stated purpose (controlling Matter via the matter CLI); nothing requested or instructed is disproportionate, though the install step downloads a binary without checksum verification — a normal but worth-noting risk.

Purpose & Capability

Name, description, and runtime instructions all describe using the Matter CLI to browse, search, tag, archive, summarize, and manage a Matter reading library. Required artifacts and commands (matter binary, matter login) match the claimed capability.

Instruction Scope

SKILL.md confines actions to Matter CLI commands and pagination/rate-limit guidance. It does not instruct reading unrelated files, accessing unrelated environment variables, or sending data to third-party endpoints outside the official GitHub releases URL or the user's Matter account via the CLI.

Install Mechanism

Install uses curl to download prebuilt binaries from the project's GitHub releases into ~/.matter/bin and makes them executable — reasonable for a CLI but lacks checksum or signature verification. Direct binary download is more risky than a package manager install because it executes code from the network without integrity checks.

Credentials

Skill declares no required env vars or credentials. It expects the user to run 'matter login' which stores auth locally — this is appropriate for a CLI-oriented Matter skill. No unrelated credentials or config paths are requested.

Persistence & Privilege

Skill is instruction-only, no install script modifies other skills or system-wide settings. always is false and the skill does not request elevated or permanent system privileges. Note: model invocation is enabled by default (agent may call the skill autonomously), which is normal for skills but worth considering because it allows automated access to the user's Matter library via the installed CLI.

Assessment

This skill appears to do what it says: it instructs you to download and run the official matter CLI and then use it to manage your Matter library. Before installing, verify the GitHub repository and the release you download (confirm repo URL is correct and review release notes). Prefer using a package manager or verifying a SHA256/GPG signature if available; if not, download only from the official getmatterapp repo and inspect the binary source if you can. Be aware 'matter login' will store credentials locally — only use accounts you trust and consider running the CLI in a restricted environment if you are cautious. Finally, consider whether you want to allow the agent to call this skill autonomously (model invocation is enabled by default) since that grants automated access to your Matter data once the CLI is installed.

Matter CLI Skill

Setup (first time)

# Install (downloads binary directly from official GitHub releases)
# macOS arm64:
curl -fsSL https://github.com/getmatterapp/matter-cli/releases/latest/download/matter-darwin-arm64 -o ~/.matter/bin/matter && chmod +x ~/.matter/bin/matter

# macOS x86:
curl -fsSL https://github.com/getmatterapp/matter-cli/releases/latest/download/matter-darwin-x64 -o ~/.matter/bin/matter && chmod +x ~/.matter/bin/matter

# Linux arm64:
curl -fsSL https://github.com/getmatterapp/matter-cli/releases/latest/download/matter-linux-arm64 -o ~/.matter/bin/matter && chmod +x ~/.matter/bin/matter

# Linux x86:
curl -fsSL https://github.com/getmatterapp/matter-cli/releases/latest/download/matter-linux-x64 -o ~/.matter/bin/matter && chmod +x ~/.matter/bin/matter

# Add to PATH if needed
export PATH="$HOME/.matter/bin:$PATH"

# Authenticate
matter login

Source: https://github.com/getmatterapp/matter-cli Requires a Matter Pro subscription. Get one at https://web.getmatter.com/settings.

CLI binary: matter (added to PATH after install) Auth: Stored locally after matter login. Full command reference: See references/commands.md

Core Workflows

Browse Inbox

matter items list --status inbox --order inbox_position --limit 10

Browse Reading Queue

matter items list --status queue --order library_position --limit 10

Read / Summarize an Article

matter items get <id> --include markdown

Fetch markdown content, then summarize or analyze it.

Search by Topic

matter search "<topic>" --type items --limit 10

Get Highlights for an Article

matter annotations list --item <item_id> --all

Tag an Article

matter tags list   # find or confirm tag id/name first
matter tags add --item <item_id> --name "<tag_name>"

Archive an Article

matter items update <id> --status archive

Save a New URL

matter items save --url "<url>" --status queue

Compile Highlights Digest

1. Search or list items by topic/tag
2. For each item: matter annotations list --item <id> --all
3. Compile highlights into structured output

Output Format

All commands return JSON by default. Use --plain for human-readable output.

⚠️ Do NOT use --all for large libraries. It streams all pages in one call and will hit rate limits mid-stream, corrupting the JSON response. Instead, paginate manually:

# First page
matter items list --status queue --order updated --limit 50 --plain

# Next page (use cursor from previous response)
matter items list --status queue --order updated --limit 50 --plain --cursor <cursor>

Repeat with each returned cursor until no cursor is shown. Add sleep 5 between pages to stay within rate limits.

Rate Limits

Operation	Limit
Read (list, get)	120/min
Search	30/min
Markdown (full text)	20/min
Write (update, tag, archive)	30/min
Save (new URLs)	10/min
Burst	5/min

Rules for bulk operations:

• Fetching full markdown for multiple articles: sleep 4 between each call (stays under 20/min)
• Running multiple searches back-to-back: sleep 3 between each (stays under 30/min)
• Never fire more than 5 requests in rapid succession (burst limit)
• If you receive a Rate limit exceeded. Retry after N seconds error: stop immediately, wait the stated seconds + 5, then resume
• For research tasks across 5+ articles: fetch markdown sequentially with delays, not in parallel