{"version":"2.0.75","anchor":"reduced-subagent-prompt-token-usage","canonical_anchor":"reduced-subagent-prompt-token-usage","heading":"Reduced Subagent Prompt Token Usage","tier":null,"area":null,"scope":null,"heads_up":null,"url":"https:\/\/changelogs.core-directive.com\/v\/2.0.75\/e\/reduced-subagent-prompt-token-usage","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.0.75","markdown":"### Reduced Subagent Prompt Token Usage\n\n**What**\n\nRemoved a duplicate instruction from subagent prompts about not using colons before tool calls.\n\n**Details**\n\n- The instruction \"Do not use a colon before tool calls. Text like 'Let me read the file:' followed by a read tool call should just be 'Let me read the file.' with a period.\" was removed from the subagent system prompt builder\n- This instruction remains in the main system prompt (`zb` function), so the guidance is still provided to the main conversation thread\n- Subagents inherit context from the parent, making this instruction redundant for them\n- This reduces token usage in subagent conversations without changing behavior\n\n**Evidence**\n\n`XbA()` at line 516791 (subagent prompt builder, removed duplicate colon instruction)"}