MCP server that monitors Slack for critical messages and interesting discussions, generating periodic briefs using Claude AI and the Henchman MCP worker
Two tools with basic schemas and descriptions, but significant gaps in parameter documentation and output schema clarity. The 'search' tool has 7 parameters with descriptions, but lacks type constraints (enums, ranges). 'getThreadDetails' has minimal parameter documentation. Neither tool documents output schema structure. Descriptions are present but generic (194-char baseline suggests room for improvement). No error handling guidance, no idempotency markers, no permission scopes declared. Tool names follow verb_noun pattern (positive), but parameter naming inconsistencies (textQuery vs query) and missing constraints reduce usability.
Retrieve full details and message content for a specific Slack thread
Search Slack threads using Henchman MCP worker with configurable filters for channels, users, text queries, and time windows
Output schemas not documented. Neither tool specifies what fields are returned, forcing LLMs to guess structure and plan downstream calls blindly.
Parameter constraints missing. 'limit' and 'since' accept unbounded values; no enums for 'since' (e.g., '2h', '24h', '48h' are examples, not constraints). LLMs will hallucinate invalid values.
Duplicate/ambiguous parameters. 'search' tool has both 'textQuery' and 'query', unclear which takes precedence or if both are required. Violates single-responsibility naming.
No error handling guidance. Tools provide no recovery hints (e.g., 'Thread not found, try search() first'). LLMs cannot self-correct on failures.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 48 | <=2025-11-25 | v2 |
Missing pagination and result limits. 'search' accepts 'limit' but no documentation of max value, default, or next_cursor for large result sets. Risks context window exhaustion.