Strip filler from internal agent communications while preserving technical content.
npx skills add https://github.com/nghiahsgs/taw-kit --skill terse-internal
Please help me install this Agent Skill. Skill: terse-internal Source: https://github.com/nghiahsgs/taw-kit/tree/main/skills/terse-internal Command: npx skills add https://github.com/nghiahsgs/taw-kit --skill terse-internal
Strip filler from Claude's English output when it's talking to itself or between agents. Keep full technical substance. Never touch user-facing Vietnamese.
Enable terse-internal for internal agent communications to trim filler while preserving technical content.
High-intent search queries and answers about installing and using this skill.
To reduce token waste in internal agent communications, strip filler like preambles, postambles, and narration from messages while preserving technical content and exact error messages. This maintains substance across planner, researcher, and debug roles.
Caveman-style terse output for internal agents is a token-optimization technique that strips conversational filler from internal messages. It removes narration and preambles while keeping code blocks, exact technical terms, and user-facing text fully intact.
Yes, you can optimize agent tokens without breaking code blocks and error messages by enforcing constraints that preserve exact technical terms. The token reduction process only removes filler and narration, ensuring technical substance remains untouched.
Terse output does not affect user-facing text during agent token optimization. The stripping process applies strictly to internal agent messages, keeping user-facing verbose content and exact technical terms fully preserved.
Token optimization for agent communications supports multiple internal roles including planner, researcher, fullstack-dev, tester, reviewer, and debug contexts. It applies uniformly across these roles to trim filler while maintaining technical content.
burtenshawOrchestrate multi-agent Autolab Hermes delegation with role contracts and parent rules.
SamuelSahaReduce AI token consumption via shallow skill loading, context pruning, and output compression.
wade56754Orchestrate multi-agent coding pipelines to generate, integrate, and verify software code.
nobe4Enforce plain language and grug principles to produce concise AI responses.
