Loading ToolBench

Reading the current corpus…

forsety MCP server · F · ToolBench

forsety

MCP server for Forsety RecallVault - provides tools for memory management, dataset access, policy checking, and semantic search with authentication, policy enforcement, and audit logging

<=2025-11-25STREAMABLE_HTTP7 tools

47/100

Definition Quality44

Protocol Readiness62

Supportability41

Spec posture

Inferred effective spec <=2025-11-25·current baseline 2026-07-28

Static source inference · medium confidence · detected: stateful session

Deprecated protocol patterns detected

Summary

Server provides 7 tools with basic schemas and descriptions, but falls short of production quality. Tool names lack clear action verbs (e.g., 'forsety_memory_store' vs 'store_memory'), descriptions are minimal (avg ~60 chars, below 194-char baseline), and parameter descriptions are sparse. No output schemas documented. Error handling and recovery guidance absent. Input schemas present but lack depth (no enums, ranges, or format constraints). Composition is reasonable (single responsibility per tool), but descriptions don't guide LLM selection or explain when to use each tool vs alternatives.

Tools (7)

forsety_dataset_accessread onlyauthsource verified40/100

Access a dataset and log the access with a cryptographic proof.

forsety_memory_deletedestructiveauthsource verified42/100

Delete a memory by namespace and key.

forsety_memory_retrieveread onlyauthsource verified43/100

Retrieve a specific memory by namespace and key.

forsety_memory_searchread onlyauthsource verified48/100

Search memories by namespace, tags, or key pattern.

forsety_memory_storewriteauthsource verified43/100

Store a memory in RecallVault. Upserts by namespace+key.

forsety_policy_checkread onlyauthsource verified45/100

Check if the agent has policy access to a dataset.

forsety_semantic_searchread onlyauthsource verified

Issues (5)

HIGHtool-naming ↗

Tool names lack action verbs. 'forsety_memory_store' should be 'store_memory' or 'upsert_memory'; 'forsety_dataset_access' should be 'access_dataset'. LLMs infer intent from verb-noun structure.

forsety_memory_storeforsety_memory_retrieveforsety_memory_searchforsety_memory_deleteforsety_dataset_accessforsety_policy_check

HIGHtool-description ↗

Descriptions are too brief (avg ~60 chars vs 194-char baseline). 'Store a memory in RecallVault. Upserts by namespace+key.' lacks context on WHEN to use this vs alternatives, what happens on retry, or what the return value contains.

forsety_memory_storeforsety_memory_retrieveforsety_memory_searchforsety_memory_deleteforsety_dataset_access

HIGHresponse-shaper ↗

No output schemas documented. LLMs cannot plan downstream calls or extract required fields. What does 'forsety_memory_retrieve' return? A string? An object with metadata? Pagination support?

forsety_memory_store

Recommendations

Rename tools to verb_noun pattern: 'store_memory', 'retrieve_memory', 'search_memories', 'delete_memory', 'access_dataset', 'check_policy', 'search_semantic'. This enables LLMs to infer intent from the name alone.
Expand descriptions to 100 - 200 chars. For each tool, answer: What does it do? When should I call it instead of a similar tool? What does it return? Example: 'Store or update a memory in RecallVault by namespace and key. Use this to persist agent state, context, or learned facts. Returns the stored value and a timestamp. Upserts silently if the key exists.'
Document output schemas for all tools. Specify return type (object, array, string), field names, and types. Example: 'Returns {value: string, namespace: string, key: string, tags: string[], created_at: ISO8601, updated_at: ISO8601}'.
Add parameter constraints and examples. For 'tags' array, specify format: 'Comma-separated or JSON array of strings, max 10 tags, each 1 - 50 chars.' For 'query', add: 'Natural language or regex pattern. Examples: "urgent", "project:*", "created:2024-*".'
Add error handling and recovery hints. Example for retrieve: 'If key not found, returns 404 with message "Memory not found. Try search_memories() with a partial key or tag." Retries are safe (idempotent).' For delete: 'Destructive operation. Returns 200 on success, 404 if already deleted (safe to retry).' For policy_check: 'Returns {allowed: boolean, reason: string}. If denied, reason explains which policy blocks access.'

Spec posture evidence

Inferred effective spec: <=2025-11-25.

Relies on Stateful initialize / Mcp-Session-Id (removed; protocol is stateless) - make each request self-contained

Score history

Overall score trend

First recorded score · v2 rubric

47/100

Scored	Grade	Overall	Spec posture	Rubric
2026-09-23	F	47	<=2025-11-25	v2