Observed arrival · 2026-09-25
Agentic Keychain measures whether an AI skill earns its price
A registry of skills for AI agents that evaluates whether a skill’s measured task savings justify buying it.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- AI agents evaluating task-specific skills
- Worth noticing
- The evaluation can return four decision states, including undetermined and no_relevant_capability.
Field notes
An agent can submit task text through an MCP evaluate call or the documented POST endpoint and receive one of four decisions: recommend, do_not_recommend, undetermined, or no_relevant_capability. The response is described as carrying measured effect, price, confidence, and an operator-facing message. Log Triage is priced at $4 one-time; Repo Cartographer is listed at $5, but its checkout is not configured. The page identifies the software as version 0.1.0.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
$PaidCommerce or pricing visible
✦PrettyNotable craft visible
●ProPolished or operationally mature
◎NicheUnusually specific use
One card from the complete issue