agent-skill vs cobbles

Side by side on the facts. GitHub figures come straight from each repository; ratings come from signed-in reviewers here.

Metricagent-skillMCP Server cobblesMCP Server

Type: agent-skillMCP Server; cobblesMCP Server
Category: agent-skillSecurity; cobblesSecurity
GitHub stars: agent-skill0; cobbles0
Trust scoreⓘAuditability and maintenance signals — not a security audit.: agent-skillNot yet assessed; cobblesNot yet assessed
Community rating: agent-skillNo reviews yet; cobblesNo reviews yet
Last commit: agent-skillUnknown; cobblesUnknown
Licence: agent-skillNone declared; cobblesNone declared
Official: agent-skillNo; cobblesNo
Source: agent-skill Repository; cobbles Repository

Highlighted cells win that row on an objective measure. Rows without a clear winner — category, licence, source — are left unmarked. A higher trust score means better auditability and maintenance, not safer code.