Skip to content
Metatext Decide — typed decisions for your agents, with a probability on every answerGet early access
Metatext
Sign Up

tink-route

CLI that routes a task to one specialist Agent Skill from a Tink library using Jev, then verifies and prints it.

Agent skillbeta
View on GitHub ↗

Overview

tink-route picks the single most suitable skill for a task from a user's Tink skill library, verifies it with tink mount, and prints the skill content (or its path for large skills). Jev makes the routing decision and returns the winner, its probability, a runner-up and a margin. If nothing fits, it says so and exits non-zero.

It is meant for coding agents: a one-line instruction in AGENTS.md tells the agent to call tink-route when it needs a specialized procedure and to continue without it if the command fails. Options include restricting routing to approved skills, writing JSON receipts for each delivery, and a --pick mode that decides without mounting anything. Delivery fails open, with a stable reason slug for every refusal.

What the decision model decides

TypeQuestionAnswers
ChoiceWhich specialist skill in the library fits this task?—
Yes / noDoes any specialist skill apply to this task?yes · no

How it works

  1. The agent runs tink-route "<task>".
  2. Jev picks a skill from the library and reports probability, runner-up and margin.
  3. If probability is below the threshold, it reports no skill and exits 1.
  4. The winner is verified with tink mount and printed inline or by path.

Strengths

  • Agents load only the one relevant skill instead of the whole library.
  • Approved-only mode prevents unapproved skills from being delivered.
  • Receipts and JSON output give an audit trail.
  • Fails open with clear reasons so the agent can continue.

Caveats

  • Requires the tink CLI and a TypeSafe API key.
  • Not published on PyPI; installed from git.

Getting started

pipx install git+https://github.com/jon-devlapaz/tink-route.git, store the TypeSafe key, then run tink-route "<task>".

At a glance

Type
Agent skill
Status
beta
Language
Python
Runtime
CLI
Has code
yes

Use cases and integrations

How curated lists describe it

Frequently asked questions

What does tink-route decide?

Which one skill in your Tink library best fits the task, or that none does.

How do agents use it?

Add one line to AGENTS.md telling the agent to run tink-route and follow its output.

What happens if Jev is unsure?

Below the threshold it reports that no specialist skill applies and exits non-zero.

Related projects

  • jev-decision-layer

    MIT plugin and local MCP server that hands an agent's closed questions to Jev or local Laya, with policy gates and receipts.

    Agent skillproductionPython
  • agent-dispatcher

    Claude Code and Codex skill that routes a task to one of 27 specialist roles, with an optional Jev engine for catalog selection.

    ProjectbetaPython
  • Gatekeeper

    Claude Code hook that decides which agent or skill should handle a request using YAML rulebooks plus one Jev Choice with confidence bands.

    HookbetaPython
  • Helm

    Orchestrator that probes which coding agents are installed and logged in, then has Jev pick the best one for a task.

    Agent skillbetaPython
  • jev-agent-hooks

    Hooks for Claude Code, Codex and pi that use Jev to suggest a skill each turn, plus a rule gate for subagent model routing.

    HookbetaJavaScript
  • jev-pilot

    Claude Code plugin that asks Jev before each turn for reasoning effort, subagent model, strategy and skill, with a decision log.

    Agent skillbetaTypeScript