Source Intelligence
Follow Discord

DisclaimerUnofficial, and not affiliated with Anthropic. Nearly all of this is read straight out of what ships: npm bundles, captured prompts, published docs. Anthropic's own notes go in verbatim, marked as theirs. The rest is my reading, and every entry carries the strings behind it. If one looks wrong, vote it down and say why.

← All of v2.1.121 Home All releases ←olderv2.1.120 v2.1.122newer→
Claude Code v2.1.121

Cost Attribution by Agent, Skill, and Plugin

What

The usage/cost display now breaks down spending by agent, skill, and plugin, showing what percentage of usage each component consumed.

Details
  • API responses now carry attributionAgent, attributionSkill, and attributionPlugin fields
  • Costs are accumulated per-request into byAgent, bySkill, and byPlugin maps
  • The cost view renders a "Skills, subagents, and plugins" section with percentage breakdown
  • Actionable tips appear alongside high-cost items:
  • Skills: "Heavy skills can be scoped down or run with a cheaper model via skill frontmatter."
  • Plugins: "Review what this plugin contributes — its agents, skills, and MCP tools all count toward your limit."
  • Shows "No attribution data yet · accumulates as you use Claude" when no data exists
Evidence

Attribution tracking (search for "attributionAgent", "attributionPlugin", "Skills, subagents, and plugins")

Strings lifted out of the shipped bundle, so the claim above can be checked against them.

See this entry in the whole of v2.1.121 →