Miniflux

Manage RSS feeds and entries on a Miniflux instance. Handles requests like "show my unread articles", "list my feeds", "rename this category", "unsubscribe f...

Security Scan

The skill is internally consistent: it is an instruction-only wrapper that expects a local MCP helper binary and a Miniflux URL + API token (or username/password) — nothing in the instructions asks for unrelated secrets or system-wide access.

Purpose & Capability

The name/description (Miniflux RSS management) matches the instructions. The only external requirement is the openclaw-miniflux-mcp helper binary and Miniflux credentials/URL, which are directly relevant to operating a Miniflux instance.

Instruction Scope

SKILL.md limits actions to Miniflux operations (list feeds, read entries, create/modify feeds/categories, import/export OPML). It does not instruct reading unrelated files or exfiltrating data to unexpected endpoints. It directs the user to configure an MCP server entry containing MINIFLUX_URL and credentials; that is expected for connecting to Miniflux.

Install Mechanism

There is no automated install spec (lower risk). The instructions tell the user to download a binary from a GitHub Releases page or install via cargo. GitHub Releases is a reasonable source but the repo is a third‑party user repo (sinhong2011); users should verify release authenticity and prefer cargo install or verified releases if possible.

Credentials

The skill requires Miniflux credentials (API token or username/password) and the instance URL to function — these are proportionate. Registry metadata lists no required env vars, but SKILL.md correctly instructs embedding MINIFLUX_URL and credentials in the MCP server config. No unrelated secrets are requested.

Persistence & Privilege

The skill is not always-enabled and is user-invocable. It does not request persistent system-wide privileges or attempt to modify other skills. Adding the MCP server is a local configuration change the user performs; that is normal for MCP-based skills.

Assessment

This skill appears to do what it says, but before installing: 1) verify the openclaw-miniflux-mcp binary source (check the GitHub Releases page, release checksums/signatures, or prefer cargo install), 2) create and use a Miniflux API key with only the needed privileges (or use read-only mode if you don't want writes), 3) store credentials in your local MCP server config (not in public/shared places), and 4) be aware that if you allow the agent to invoke skills autonomously it may perform Miniflux actions without an interactive prompt — enable autonomous use only if you trust the agent's behavior. If you want extra assurance, request the binary's source code or a signed release from the maintainer before downloading.

What it does

Provides access to a Miniflux RSS reader instance through 13 read tools and 10 write tools. Agents can browse feeds, search entries by status or date, read specific articles, check categories, and (if not in read-only mode) create/update/delete feeds and categories, import OPML, mark entries as read, and toggle bookmarks.

Inputs needed

• For listing entries: status, date range, starred, pagination (all optional)
• For feed-specific queries: feed ID
• For single items: entry ID, feed ID, or user ID
• For discovery: a URL to scan for feeds
• For creating categories: a title
• For updating categories: category ID and new title
• For deleting categories: category ID
• For subscribing to feeds: a feed URL and category ID
• For updating feeds: feed ID, plus optional title, category_id, feed_url, site_url, user_agent
• For deleting feeds: feed ID
• For OPML import: an OPML XML string
• For writes: entry IDs + status, or entry ID for bookmark toggle

Prerequisites

openclaw-miniflux-mcp binary

Download the latest binary for your platform from GitHub Releases:

Platform	Binary
Linux x86_64	openclaw-miniflux-mcp-x86_64-unknown-linux-gnu
Linux ARM64	openclaw-miniflux-mcp-aarch64-unknown-linux-gnu
macOS x86_64	openclaw-miniflux-mcp-x86_64-apple-darwin
macOS ARM64	openclaw-miniflux-mcp-aarch64-apple-darwin

Or install via Cargo:

cargo install openclaw-miniflux-mcp

MCP server configuration

Add the MCP server to your client configuration:

With API token (recommended):

{
  "mcpServers": {
    "miniflux": {
      "command": "/path/to/openclaw-miniflux-mcp",
      "args": [],
      "env": {
        "MINIFLUX_URL": "http://localhost:8080",
        "MINIFLUX_API_TOKEN": "<your-api-token>"
      }
    }
  }
}

With username/password:

{
  "mcpServers": {
    "miniflux": {
      "command": "/path/to/openclaw-miniflux-mcp",
      "args": [],
      "env": {
        "MINIFLUX_URL": "http://localhost:8080",
        "MINIFLUX_USERNAME": "<username>",
        "MINIFLUX_PASSWORD": "<password>"
      }
    }
  }
}

Read-only mode (disables write tools):

{
  "mcpServers": {
    "miniflux": {
      "command": "/path/to/openclaw-miniflux-mcp",
      "args": ["--read-only"],
      "env": {
        "MINIFLUX_URL": "http://localhost:8080",
        "MINIFLUX_API_TOKEN": "<your-api-token>"
      }
    }
  }
}

The user will need to:

1. Replace the binary path with wherever they downloaded/installed it
2. Replace MINIFLUX_URL with their Miniflux instance URL
3. Get an API token from Miniflux: Settings > API Keys > Create a new API key
4. Restart their MCP client after saving

Workflow

Browsing feeds

1. Call miniflux_get_feeds to see all subscriptions
2. Call miniflux_get_feed_entries with a feed ID to see its entries
3. Call miniflux_get_entry to read a specific article

Searching entries

Call miniflux_get_entries with filters:

• status: "unread", "read", or "removed"
• starred: true for bookmarked entries
• after / before: Unix timestamps for date ranges
• limit: Number of results (default varies, recommend 20)
• order: "published_at" and direction: "desc" for newest first

Triaging unread articles

1. Call miniflux_get_entries with status: "unread", limit: 20
2. Read interesting entries with miniflux_get_entry
3. Mark reviewed entries as read: miniflux_update_entry_status with status: "read"
4. Bookmark important ones: miniflux_toggle_bookmark

Adding new feeds

1. Call miniflux_discover_subscription with a website URL to find available feeds
2. Present discovered feeds to the user
3. If needed, call miniflux_create_category to create a new category
4. Call miniflux_create_feed with the feed URL and category ID to subscribe

Managing feeds

• Update: miniflux_update_feed with feed ID and any fields to change (title, category_id, feed_url, site_url, user_agent)
• Delete: miniflux_delete_feed with feed ID to unsubscribe
• Refresh: miniflux_refresh_feed with feed ID to fetch new entries now

Managing categories

• List: miniflux_get_categories to see all categories
• Create: miniflux_create_category with a title
• Rename: miniflux_update_category with category ID and new title
• Delete: miniflux_delete_category with category ID (feeds move to default category)

Importing/Exporting

• Export: miniflux_export_opml to get all feeds as OPML XML
• Import: miniflux_import_opml with an OPML XML string to bulk-add feeds

Guardrails

• Default to small page sizes (limit=20) to avoid overwhelming responses
• On 401/403 errors, tell the user to check their API token or credentials
• On connection errors, tell the user to verify their MINIFLUX_URL
• Confirm with the user before marking large batches of entries as read
• In read-only mode, explain the limitation clearly when a write is attempted
• When listing returns empty results, suggest checking filters or confirming the instance has data