An MCP server that exposes prompt decorators as MCP tools, allowing any MCP client to apply various prompt enhancement decorators to improve interactions with LLMs
The Prompt Decorators MCP server has 7 tools covering decorator discovery, application, and validation. Tool naming is generally good (verb-first pattern: list_, get_, apply_, validate_, search_), and descriptions are present for all tools and parameters. However, there are significant gaps in schema completeness and parameter documentation. Most tools have descriptions in the 40-100 character range, which is brief but adequate. Input schemas are visible in the code and include type information (string, object, array), but lack detailed constraints (enums, min/max values, patterns). Parameter descriptions are minimal and don't always explain the expected format or valid ranges. Output schemas are not documented, critical for LLMs to understand what fields to expect from results. Error handling and recovery guidance are absent from all tool definitions. The server operates on READ_ONLY operations (no destructive actions), which reduces security risk but also means no idempotency, confirmation, or permission-gate patterns are tested.
Apply a prompt decorator to a given prompt text.
Apply multiple prompt decorators to a given prompt text in sequence.
Get detailed information about a specific decorator.
Get all decorators in a specific category.
Lists all available prompt decorators.
Search for decorators by name or description.
Validate the syntax of a decorator definition or decorator application syntax.
Output schemas are completely undocumented. LLMs cannot infer what fields to expect from any tool result. This violates the foundational pattern that tools must return structured, documented data.
Parameter descriptions lack specificity. For example, 'parameters' in apply_decorator says 'Optional parameters for the decorator as key-value pairs' but does not explain what keys are valid, what types of values are expected, or any constraints. The 'decorators' array in apply_multiple_decorators lacks item schema description.
No enumeration constraints on parameters. 'category' in get_decorators_by_category accepts a free-form string with no documented valid values. 'query' in search_decorators has no guidance on syntax or expected format. This invites hallucinated invalid inputs.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | D | 50 | <=2025-11-25 | v2 |
| 2026-03-09 | F | 31 | - | v1 |
No error handling or recovery guidance documented. If a tool fails (e.g., decorator not found, syntax invalid, parameters malformed), the descriptions do not hint at what the LLM should try next. This breaks the recovery-guide pattern.
Missing pagination/result-limiting guidance. list_decorators and search_decorators do not document whether they return all results or enforce a limit, and there is no mention of pagination parameters (limit, offset, next_cursor). Large decorator registries could exhaust context windows.