verification-before-completion 95 /100 Evidence requirement enforcement ensuring all claims are backed by logs, test results, or exit codes. Zero = success, non-zero = failure. No guessing allowed.
uses shell commands
Type: Agent skill (SKILL.md) · Category: Testing & QA · Source: https://github.com/a5c-ai/babysitter/tree/HEAD/library/methodologies/cc10x/skill
Score breakdown Capabilities (fewer powerful permissions = higher) 45/50 Maintenance 35/35 Documentation 15/15
Why this score Runs shell commands or spawns processes (instructions/allowed tools ask the agent to) −5 Instructions only: no bundled scripts Last updated 29 days ago +15 1,847 GitHub stars +10 42 contributors +5 License: MIT +5 Description and source link present +15 Always review files before installing. How scores work . Request a correction
Weekly: new skills & MCP servers, scored.
Scores are automated static checks of public files and metadata, not a security audit. They describe what an item can do, not its authors' intent. Updated 2026-10-16.