Skill invocation fails with '1M context extra usage required' even when already on 1M model

Resolved 💬 3 comments Opened Apr 12, 2026 by lisa3907 Closed May 24, 2026

Description

When invoking a custom skill via /pds-test (a Playwright MCP-based E2E integration test skill), the following error is consistently thrown:

API Error: Extra usage is required for 1M context · run /extra-usage to enable, or /model to switch to standard context

Environment

  • Claude Code version: Latest (CLI)
  • Model: claude-opus-4-6[1m] (already on 1M context)
  • OS: Windows 11 Pro 10.0.26200
  • Shell: bash (Git Bash)

Steps to reproduce

  1. Start Claude Code with Opus 4.6 (1M context) model
  2. Have a custom skill defined in .claude/skills/pds-test/
  3. Run /pds-test in the conversation

Expected behavior

The skill should load and execute normally, since the session is already running on the 1M context model.

Actual behavior

The error API Error: Extra usage is required for 1M context is thrown immediately when the skill is invoked. The skill file itself is ~200 lines of markdown instructions with references to other files — it should not trigger a context limit issue on invocation alone.

Notes

  • The skill works conceptually — it reads reference files and drives Playwright MCP tools
  • The error appears to be a billing/entitlement gate, not an actual context overflow
  • Running /extra-usage is not always an option depending on the plan
  • The error message suggests switching models, but the skill requires 1M context to function properly with Playwright MCP snapshots

View original on GitHub ↗

This issue has 3 comments on GitHub. Read the full discussion on GitHub ↗