Loading ToolBench
Token-efficient AI coding setup CLI (aito) — installs and configures a curated toolkit of components for reducing token usage in AI-assisted development workflows
This bash CLI exposes 7 tools via STDIO-only MCP transport. Critical gaps: NO input schemas visible for ANY tool except 'learn' (which has minimal schema). Most tool descriptions are under 50 chars and lack actionable detail. No output schemas documented. No error handling guidance. Tool names lack clear verb-noun structure (e.g., 'setup' is ambiguous, setup what? for whom?). Parameters are not formally typed or constrained. This server reads like a CLI wrapper, not a production-grade agent tool interface.
Check config files, token budgets, and tool availability
Print detected environment
Show help information
Append a deduped lesson to this project's playbook.md (ACE-style)
Interactively install/configure tooling in the current project
Measure token reduction; write token-report.md
Print version
NO input schemas visible for 6 of 7 tools. Only 'learn' has a minimal schema with one string parameter. Tools like 'setup', 'verify', 'doctor' accept no documented parameters, making it impossible for LLMs to understand what inputs they accept or require.
Descriptions are too short and lack actionable context. 'Measure token reduction; write token-report.md' (verify) and 'Check config files, token budgets, and tool availability' (doctor) are under 50 chars and do not explain WHEN to call these tools, WHAT they return, or HOW to interpret results. LLMs cannot reliably select these tools.
Tool names lack clear verb-noun structure. 'setup', 'verify', 'doctor', 'learn', 'env' are vague. 'setup' could mean setup project, setup tools, setup config, ambiguous. 'doctor' is a metaphor, not an action verb. 'env' is truncated. Recommend: 'configure_project_tools', 'measure_token_reduction', 'check_project_health', 'append_lesson_to_playbook', 'print_environment'.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 30 | 2026-07-28+ | v2 |
No output schemas documented. Tools like 'verify' write 'token-report.md' but the MCP schema does not declare what fields or structure the response contains. LLMs cannot plan downstream actions or extract specific data.
No error handling guidance. If 'setup' fails (e.g., git not found, component installer crashes), the tool warns but does not tell the LLM what to do next. No recovery paths, no retryable vs. fatal classification, no actionable error messages.