Loading ToolBench
A Next.js-based mission control and agent orchestration platform with integrations for approvals, documents, tasks, calendar, and agent management
Arkon exposes 21 tools with basic naming and descriptions, but critical gaps undermine production readiness. Tool names follow verb_noun convention (approvals_list, tasks_update_status), but descriptions are often generic and lack actionable context. Input schemas are present but incomplete: many parameters lack type constraints (enums for status fields), and output schemas are entirely undocumented. No error handling guidance, no recovery paths, and no indication of which tools are idempotent or destructive. Security-sensitive tools (admin_agents_create, admin_budgets_delete) lack permission documentation. Pagination is inconsistent, some tools accept offset/limit, others do not. Parameter descriptions are minimal (10-30 chars), below the 72-char baseline for A+ tools.
Get list of active agent runs and live agents
Create a new agent with optional tenant and provisioning
List all agents with role and tenant information
Update agent role, tenant, or rotate token
Create a new budget limit for tenant or agent
Delete a budget limit
List all budget limits with current spend tracking
Update an existing budget limit
Output schemas completely undocumented. No tool declares what fields it returns, forcing LLMs to guess structure and plan downstream calls blindly.
Status and action parameters lack enum constraints. 'status' accepts free-form strings (e.g., 'pending', 'approved', 'rejected') with no validation guidance, inviting hallucinated values.
Destructive tools (admin_budgets_delete, admin_agents_update with rotate_token) lack confirmation/dry-run patterns and permission documentation. No guidance on recovery if called in error.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 48 | 2026-07-28+ | v2 |
List all synced cron jobs
Sync cron jobs from gateway or send message about a cron job
Enable, disable, delete, or run a cron job
Fetch approval items with optional status filtering and pagination
Update the status of an approval item (approve, reject, or expire)
Fetch calendar items with optional date range filtering
Update the status of a calendar item
Fetch document items with optional category filtering and search
Pin or unpin a document for quick access
Fetch quick command history
Fetch subagent run history with optional filtering
Fetch task items with optional status and priority filtering
Update the status of a task item
Parameter descriptions are minimal (10-40 chars, well below 72-char baseline). E.g., 'Filter by approval status' lacks format, range, or dependency hints. LLMs cannot infer valid values.
No error handling guidance. Tools do not document what errors are retryable, user-fixable, or fatal. No recovery paths (e.g., 'If agent not found, call admin_agents_list first').