Optimize prompt-cache hits and strip redundant context from AI harness configurations.
npx skills add https://github.com/AesopScott/mojo --skill cache-optimizer
Please help me install this Agent Skill. Skill: cache-optimizer Source: https://github.com/AesopScott/mojo/tree/main/harnesses/skills/cache-optimizer Command: npx skills add https://github.com/AesopScott/mojo --skill cache-optimizer
This Skill addresses excessive AI operational costs and token waste by optimizing prompt-cache utilization and removing redundant context from instruction payloads.
Run the cache-optimizer skill to analyze the current Prompt harness and identify redundant context that can be removed to improve cache efficiency.
High-intent search queries and answers about installing and using this skill.
Reducing input token consumption requires maximizing prompt-cache hits and stripping redundant context from system instruction payloads. Auditing injection sequences and removing unnecessary data lowers the token count per API call.
Prompt-cache optimization refines prompt templates and injection order to ensure maximum reuse of cached prompt segments. Systematically inspecting harness configurations and removing redundant context directly lowers the cost per API call.
Auditing a prompt harness for redundant system instructions involves systematically inspecting current configurations and verifying token usage metrics. This process identifies and prunes unnecessary data in injection sequences to maximize prompt-cache hits.
Yes, cache-optimizer is designed for high-volume prompt harnesses. It audits and prunes redundant system instructions from the instruction payload, significantly lowering the input token consumption and cost per API call.
The best way to maximize prompt-cache hits is to refine templates and adjust the injection sequence order to ensure maximum reuse of cached prompt segments. This requires systematic inspection of the current harness configurations.
Prompt-cache optimization is not ideal when injection sequences contain highly dynamic, non-repeating context that cannot be cached. The process relies on identifying and stripping redundant data from stable templates to reduce input token consumption.
dkyazzentwatwaGenerate session efficiency reports with token savings and cache performance metrics.
barkleesandersScan Claude Code JSONL session files and generate markdown token usage reports.
AesopScottManage prompt template libraries and injection-order rules for AI harnesses.
markus41Implement prompt caching for Claude API interactions with configurable TTL options.
