Vendor-neutral personal-assistant plugin: AI digest, daily briefing, personal branding, LinkedIn toolkit, newsletters, writing quality, and visual artifacts. Works with Claude Code, GitHub Copilot, and Codex.
This server has significant definition quality gaps. Tool names follow verb_noun convention (fetch_sources, run_digest, memory_remember), which is good. However, descriptions vary widely in quality, some are detailed (fetch_sources: 180+ chars), others are vague (newsletter_stats, newsletter_init_db: <50 chars). Most critically, input schemas are present but parameter descriptions are often generic or missing context. For example, newsletter_log_event's 'meta' parameter is described only as 'Optional metadata as JSON' with no guidance on structure. Error handling is absent, no recovery guidance, no categorization of retryable vs fatal errors. Output schemas are not documented anywhere. The server lacks idempotency hints, permission gates, and audit trail patterns. No tool declares what it modifies or its side effects clearly.
Reads the `digest.sources` list from the assistant config, fetches recent content from RSS feeds and web pages, and saves structured JSON for the digest skill to process.
Append a running note to notes.md
Print the resolved file path for memory or notes
Append a durable profile fact to memory.md
Print a memory or notes file
Add a subscriber to the newsletter database
Export subscribers to email-list.json
Deploy newsletter HTML and tracker.php to FTP server
No output schemas documented. Tools return data but LLMs cannot predict field names, types, or structure. Downstream tool chaining is impossible without trial-and-error.
Vague parameter descriptions. 'meta' in newsletter_log_event is 'Optional metadata as JSON' with no structure guidance. 'heading' in memory_remember lacks context on valid values or format.
No error handling or recovery guidance. Tools fail silently or return raw errors. No indication of retryability, user-fixable vs fatal, or next steps for LLM.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 42 | 2026-07-28+ | v2 |
Generate blank newsletter HTML shell or inject per-subscriber tracking
Import subscribers from email-list.json
Create or migrate the newsletter SQLite database schema
List all newsletter subscribers
Log a newsletter event (open, click, bounce)
Recalculate funnel stages for all active subscribers based on engagement metrics
Register a new newsletter in the database
Remove a subscriber from the newsletter database
Show newsletter dashboard statistics
Reads the `profile` section of the assistant config and reports which fields are incomplete.
Invokes the /ai-digest skill via the assistant CLI and optionally opens the result in a browser. Use this for scheduling or manual one-click runs.
Destructive operations lack confirmation or dry-run. newsletter_remove_subscriber and newsletter_recalc_funnel modify state with no undo path or confirmation step.
No pagination or result limits documented. fetch_sources accepts max_items but no guidance on default, max cap, or how to handle large result sets.