Enforces worker-tier and agent-count caps in sub-agent dispatch harnesses.
npx skills add https://github.com/AesopScott/mojo --skill worker-ceiling
Please help me install this Agent Skill. Skill: worker-ceiling Source: https://github.com/AesopScott/mojo/tree/main/harnesses/skills/worker-ceiling Command: npx skills add https://github.com/AesopScott/mojo --skill worker-ceiling
This Skill addresses runaway costs and resource waste in sub-agent dispatch systems by providing a structured framework to enforce operational boundaries.
Use the worker-ceiling skill to inspect the current sub-agent dispatch configuration and propose a cost-limiting patch plan.
High-intent search queries and answers about installing and using this skill.
To control sub-agent dispatch costs, you enforce per-stage worker-model tiering and agent-count caps, bounding compute and token usage within the dispatch harness to prevent excessive spending.
The best way to limit token usage is by applying a worker-model tier ceiling to the dispatch process, immediately bounding spend and enforcing strict operational safety boundaries across stages.
Early-exit logic improves dispatch efficiency by detecting empty stages and bypassing unnecessary processing, minimizing wasted compute and ensuring the sub-agent harness only executes when required.
Your sub-agent dispatch process exceeds the compute budget because it lacks strict control levers, requiring defined worker-model tiers and agent-count ceilings to bound the spend and verify execution.
Yes, you can apply agent-count caps to specific stages, enforcing resource constraints and operational safety by defining control levers that limit the number of agents executing per stage.
You need a worker-ceiling framework when a sub-agent dispatch process is exceeding budget, requiring strict control over worker-model tiers and bounded compute to prevent runaway operational costs.
AesopScottOptimizes sub-agent dispatch performance by tuning lane concurrency and fan-out parameters.
AesopScottEnforce tool-call ceilings and dry-run protocols for Tool Registry operations.
AesopScottRegister agent types and enforce result schemas in the Sub-agent Dispatch harness.
Divith123Classify AI Constitution violations and assign punishment tiers via Python scripts.
