{"version":"2.1.323","anchor":"the-skill-tool-description-in-the-compaction-prompt-was-rewr","canonical_anchor":"the-skill-tool-description-in-the-compaction-prompt-was-rewr","heading":"Skill instructions used while summarising a conversation were reworded","tier":"internal","area":"Compaction","scope":"individual","heads_up":false,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.323\/e\/the-skill-tool-description-in-the-compaction-prompt-was-rewr","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.323","markdown":"### Skill instructions used while summarising a conversation were reworded\n\nThe Skill tool's description used when Claude Code compacts a conversation was reworded, including how background skills report back\n\n**What**\n\nWhen a conversation grows too long, Claude Code compacts it, which means summarising earlier parts to free up space. The prompt used for that step carries its own copy of the Skill tool's description, the text Claude reads about loading skills. That copy was reworded in this release.\n\nIt says that a skill running in the background returns only the helper agent's name. Its result arrives later as a task notification, so Claude should not wait on it or call it again in the meantime.\n\n**Why**\n\nChanging Claude's instructions can change how it behaves. This text guides Claude away from waiting on or repeating a background skill.\n\n- Area: Compaction\n- Tier: Under the hood\n- Useful: 1\/5\n- Signal: 0\/5\n- Scope: individual\n- Heads-up: no"}