Releases·Tool Release
Claude Code subagents cut context ~88% with omitClaudeMd flag
A Reddit user found that Claude Code subagents load the project CLAUDE.md by default, inflating context use. Setting `omitClaudeMd: true` per subagent cut input tokens from 226,340 to 26,480 across 10 subagents. The flag is available in Claude Code 2.1.271+.
17 engagement1 sourceThu, Sep 17, 2026, 01:46 AM
AI Overview
- Reddit user found 10 Claude Code subagents each loaded project CLAUDE.md; setting omitClaudeMd: true cut input tokens from 226,340 to 26,480 (~88%), available in Claude Code 2.1.271+
Entities
Related
Social·Community
Tips and tricks to reduce token usage
A Reddit user posted practical techniques for cutting token consumption when working with Claude Code, including a codebase indexing tool, prompting-mode strategies, and subagent offloading. The thread invites further community contributions on token-saving workflows.
A Reddit user posted practical techniques for cutting token consumption when working with Claude Code, including a codebase indexing tool, prompting-mode strategies, and subagent offloading. The thread invites further community contributions on token-saving workflows.
12 engagement1 sourceredditThu, Sep 24, 2026, 08:16 AM
Social·Community
Users report saving 12k+ context tokens by disabling Artifacts and Chrome MCP in Claude
A Reddit user investigating degraded Claude performance found that disabling unused features like Artifacts and Chrome MCP server can save tens of thousands of startup tokens. The user reports specific token savings from these settings changes, though it did not resolve the underlying quality degradation.
A Reddit user investigating degraded Claude performance found that disabling unused features like Artifacts and Chrome MCP server can save tens of thousands of startup tokens. The user reports specific token savings from these settings changes, though it did not resolve the underlying quality degradation.
30 engagement1 sourceredditSun, Aug 23, 2026, 07:06 AM
Social·Community
Users report token savings by disabling unused Claude Code system tools
A developer reports saving 3.6K tokens per session by disabling the unused 'Artifact' system tool in Claude Code, highlighting that Anthropic bundles large tool descriptions that bloat context. The tip suggests users can reduce token usage and keep context focused by disabling non-essential system tooling.
A developer reports saving 3.6K tokens per session by disabling the unused 'Artifact' system tool in Claude Code, highlighting that Anthropic bundles large tool descriptions that bloat context. The tip suggests users can reduce token usage and keep context focused by disabling non-essential system tooling.
7 engagement1 sourceredditMon, Aug 10, 2026, 05:12 PM
Social·Community
Developer tests 3 Claude Code plugins to cut multi-agent token costs
A developer building an internal CRM with a multi-agent Claude Code setup tested three plugins aimed at reducing token costs. Analysis showed cache reads, not generated code, drove most token spend, since each turn resends growing context across agents. The post shares which cost-reduction approaches actually worked for practitioners running similar agent loops.
A developer building an internal CRM with a multi-agent Claude Code setup tested three plugins aimed at reducing token costs. Analysis showed cache reads, not generated code, drove most token spend, since each turn resends growing context across agents. The post shares which cost-reduction approaches actually worked for practitioners running similar agent loops.
27 engagement1 sourceredditMon, Aug 31, 2026, 03:57 PM
Social·Community
Stop writing handoff docs. /rewind is the tool 90% of Claude Code users have never opened.
A Reddit post argues Claude Code users mismanage context by writing handoff docs instead of trimming it. The author details a workflow pairing /rewind, /compact, and /clear with structured spec files, claiming reductions from six-figure token counts to ~30k without losing findings.
A Reddit post argues Claude Code users mismanage context by writing handoff docs instead of trimming it. The author details a workflow pairing /rewind, /compact, and /clear with structured spec files, claiming reductions from six-figure token counts to ~30k without losing findings.
13 engagement1 sourceredditMon, Sep 14, 2026, 09:43 PM