cyberneticlibrary

Render code review context

code-review-contextskillsetup L10
joeblackwaslike/agent-skills
What it does

Manage visible model context for code review

Best for

Prevent LLM model context bloat and cache misses by capping injected code-review history.

Inputs
  • · model context history
Outputs
  • · bounded context fragments with token caps
Preconditions

Context management system in place; all fragments implement ContextualUserFragment trait

Failure modes
  • · Unbounded context growth causes cache misses
  • · Items >10K tokens cross size limits
Trust signals
  • · Hard token cap per fragment
  • · Struct-based context definition
  • · P0 flagging for >1K token items