PulseAugur
EN
LIVE 04:10:56

Claude Code's skill description budget revealed to be 1% of context window

By PulseAugur Editorial · [1 sources] ·

A recent experiment with Claude Code 2.1.289 and the Opus 5.5 model revealed that the system's skill description budget is 1% of the context window, not the 2% previously stated in documentation. When this budget is exceeded, Claude Code drops descriptions of less frequently invoked skills, impacting the model's ability to autonomously select them. The experiment found that while skill names are retained, the model failed to invoke skills that only had their names listed, contrasting with its success rate for skills with descriptions. AI

IMPACT Clarifies limitations in Claude Code's skill invocation, potentially affecting developer workflows and model autonomy.

RANK_REASON Analysis of a specific product's behavior and documentation discrepancies.

Read on dev.to — Claude Code tag →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

Claude Code's skill description budget revealed to be 1% of context window

How we ranked this

Signal score
33 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
Analysis of a specific product's behavior and documentation discrepancies.
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
product, model release
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

Full methodology in our editorial standards.

COVERAGE [1]

  1. dev.to — Claude Code tag TIER_1 English(EN) · Rulestack ·

    With 160 skills installed, Claude Code kept 20 descriptions, and the model found 0 of 3 skills that lost theirs

    <blockquote> <p>With Opus 5.5's 1M-token window, Claude Code 2.1.289 held the skill listing to 30,000 characters. Of 160 project skills with 1,000-character descriptions, 20 kept their descriptions (22 of 40, 21 of 80) and every name stayed, but the model invoked 0 of 3 target sk…

RELATED ENTITIES

RELATED TOPICS