Prune context irrelevant to the active objective within a goal-oriented harness.
npx skills add https://github.com/AesopScott/mojo --skill context-pruner
Please help me install this Agent Skill. Skill: context-pruner Source: https://github.com/AesopScott/mojo/tree/main/harnesses/skills/context-pruner Command: npx skills add https://github.com/AesopScott/mojo --skill context-pruner
This Skill addresses excessive token consumption and compute waste by identifying and removing context that is not relevant to the active objective.
Invoke the context-pruner skill to analyze the current goal and remove all context that does not directly support the active objective.
High-intent search queries and answers about installing and using this skill.
Reduce token consumption by pruning irrelevant context. This Skill identifies and removes outdated logs and irrelevant conversation threads, ensuring the model focuses only on the active objective to cut compute waste.
Context boundary control dynamically prunes irrelevant data to maintain focus on the current goal. You need it in complex development workflows requiring strict boundary control over token usage and compute resources.
Prune irrelevant context by invoking this Skill to analyze the current goal and remove all data that does not directly support the active objective within your specific harness boundary.
Yes, you need to define a specific harness boundary. Safe and effective pruning requires a clear definition of both the current objective and the harness boundary to prevent removing relevant data.
Optimize token efficiency by filtering out noise before model execution. This approach strips away irrelevant context, reducing token spend and processing overhead while maintaining active goal focus.
Avoid automated context pruning when you cannot clearly define the active objective or the specific harness boundary. Without these definitions, pruning risks removing relevant data needed for safe execution.
smarmen86Restricts file edits and writes to a specified directory via bookmarking.
Ced3-hanAutomate 4-bit activation-aware quantization for large language models.
AesopScottRegister agent types and enforce result schemas in the Sub-agent Dispatch harness.
anserIndicusFilter community data by removing off-topic content, spam, and bot responses.
