Loading ToolBench
Know exactly which skills to invest in to become irreplaceable in the age of AI. 88 roles, 1,028 skills, each rated for AI impact. Python SDK + MCP server + CLI.
skillich exposes 6 well-intentioned tools with clear domain purpose (career skills taxonomy). Tool names follow verb_noun convention (explore_roles, analyze_ai_impact, assess_skills, generate_learning_path, advise_career, action_plan). Descriptions are substantive (100 - 250 chars) and explain WHEN to use each tool. Input schemas are present with type declarations and parameter descriptions. However, output schemas are NOT documented, LLMs cannot predict response structure for downstream chaining. Parameter descriptions lack format constraints (e.g., 'level' accepts 'junior|mid|senior|staff|principal' but this is stated in prose, not as enum). No error handling guidance visible. Tool descriptions are good but could be more concise (some exceed 200 chars). assess_skills and generate_learning_path accept JSON-as-string parameters (assessments, current_skills) which are error-prone, should be structured objects or arrays.
Get your personalised AI action plan: what to invest in, partner on, and delegate. Take a plain job title and return the full invest / partner / delegate breakdown in one call. This is the fastest way to understand AI impact on your specific role.
Analyze career transitions between tech roles. Find which roles you can transition to based on skill overlap, discover which of your skills are most transferable, and compare the skill gap between any two roles. Use this to make data-driven career decisions.
Analyze how AI is transforming tech roles and skills. Shows which skills are being automated, augmented, or preserved. Every skill has an AI impact rating from 0.0 (fully human) to 1.0 (fully automatable). Use this to make informed career investment decisions.
Assess your skills against a target role and identify gaps. Provide your current proficiency level for each skill and get a gap analysis showing where you're behind, on track, or ahead. Gaps are prioritized by AI resilience -- skills that AI can't easily replace are flagged as higher priority.
Explore the skillich taxonomy of tech roles and skills. Start here to browse functions, roles, skills, and proficiency levels. The taxonomy covers Engineering, Product, Design, Data & Analytics, Marketing & Growth, Operations, and Leadership.
Output schemas not documented. LLMs cannot predict response structure, breaking tool chaining and forcing exploratory calls.
assess_skills and generate_learning_path accept 'assessments' and 'current_skills' as JSON-as-string parameters. This is error-prone, LLMs must manually construct JSON strings. Should use structured array/object types.
Parameter 'level' and 'target_level' accept free-form strings but only 5 values are valid (junior|mid|senior|staff|principal). Should be declared as enum, not prose description.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 57 | 2026-07-28+ | v2 |
Generate a prioritized learning path for reaching a target role and level. Skills are prioritized by a formula: 60% AI resilience + 40% level relevance. This means skills that AI can't replace AND that matter most at your target level come first. The path is organized into growth phases.
No error handling guidance visible in tool definitions. If a role_id or function_id is invalid, LLMs have no recovery path.
Tool descriptions exceed 200 chars in several cases (analyze_ai_impact, generate_learning_path). Baseline for A+ tools is 50 - 200 chars. Trim to essential context.