Fetch, parse, search, monitor, and extract RSS/Atom feeds via @missionsquad/mcp-rss — 6 tools

@missionsquad/mcp-rss v1.0.8 — Verified Recipe

Install: npm install @missionsquad/mcp-rss Entry: dist/index.js (ESM, type: "module") Transport: stdio Auth: none required

6 Tools

Key Gotchas

1. ⚠️ Boolean params are STRINGS not booleans — useDescriptionAsContent: "true", parallel: "true", includeMetadata: "true". Passing actual boolean true causes -32602 Invalid parameters error
2. `searchIn` is a SINGLE string enum — not an array. Use "title" or "all", NOT ["title", "content"]
3. `since` accepts TWO types — either a number (Unix timestamp in milliseconds) or the literal string "last" (uses server-side cache of last fetch time). ISO date strings are NOT accepted
4. Built-in caching — after first fetch, subsequent calls to the same feed URL use cache. monitor_feed_updates with since: "last" leverages this to show only new items
5. Atom feeds work — tested with GitHub releases Atom feed (/releases.atom), parsed correctly with title, content, dates
6. Invalid URLs return text error (not MCP error) — graceful degradation with "getaddrinfo ENOTFOUND" message
7. `fetch_multiple_feeds` returns per-feed status — {total, successful, failed, results: [{url, success, data}]} so partial failures don't kill the batch
8. `extract_feed_content` markdown format — wraps each item as # Title\n\nContent with metadata when includeMetadata: "true"
9. `get_feed_headlines` markdown format — returns ### [Title](url)\nSummary per item
10. First fetch is SLOW — 7-47 seconds depending on feed size/network. Subsequent calls use cache (15-60ms)
11. `search_feed_items` only searches cached feeds — must fetch_rss_feed or fetch_multiple_feeds first, then search uses cached data (sub-100ms)
12. Server logs to stderr — clean MCP protocol on stdout

Verified Execution Trace (22 calls across 2 sessions, 16 OK + 6 param-type errors fixed)

Session 1 — initial exploration (discovered param-type gotchas):

✓ fetch HN frontpage               → 20 items, structured JSON       (14147ms)
✗ fetch CF blog (boolean true)      → -32602 Invalid parameters       (91ms)
✗ fetch_multiple (boolean parallel) → -32602 Invalid parameters       (26ms)
✓ headlines HN (json format)        → 20 headlines with summaries     (28ms)
✓ headlines HN (markdown)           → ### [Title](url) format         (15ms)
✗ search (array searchIn)           → -32602 Invalid parameters       (25ms)
✗ monitor (ISO date string)         → -32602 Invalid parameters       (136ms)
✗ extract (boolean includeMetadata) → -32602 Invalid parameters       (13ms)
✓ extract HN (json)                 → 20 items with content           (17ms)
✓ fetch GitHub Atom feed            → 10 claude-code releases         (7837ms)
✓ fetch invalid URL                 → graceful ENOTFOUND error        (267ms)
✗ search multi (array searchIn)     → -32602 Invalid parameters       (15ms)

Session 2 — all params corrected:

✓ fetch CF blog (string "true")     → 20 items, descriptions as content (47251ms)
✓ fetch_multiple (string parallel)  → 2