Update Thompson sampling parameters and track skill quality after WinDAG runs.
npx skills add https://github.com/curiositech/port-daddy --skill windags-curator
Please help me install this Agent Skill. Skill: windags-curator Source: https://github.com/curiositech/port-daddy/tree/main/skills/windags-curator Command: npx skills add https://github.com/curiositech/port-daddy --skill windags-curator
Post-execution crystallization and learning updates for WinDAGs, enabling continuous improvement of skill quality and coordination across evaluations.
After a successful DAG run, trigger the curator to apply learning updates and generate a LearningResult for the Knowledge Library.
High-intent search queries and answers about installing and using this skill.
Thompson sampling parameters are updated after a DAG run by triggering the post-execution curator to adjust skill and method parameters based on node evaluation signals.
Monster-barring detection is a post-execution learning process that surfaces fragile patterns in skill quality to avert degradation and prevent strategy failures across DAG evaluations.
Kuhnian crisis signaling identifies paradigm shifts during post-execution learning updates, enabling the crystallization of new strategies and skills within the Knowledge Library.
Near-miss conditions are logged automatically during post-execution learning updates to track skill quality and detect fragile patterns across the WinDAGs evaluation nodes.
Yes, evaluator-driven updates are applied directly to the Knowledge Library without external dependencies, generating a LearningResult that tracks methods and updates learning signals.
Post-execution crystallization should be triggered after a successful DAG run to apply learning updates, detect crisis conditions, and crystallize new strategies for continuous improvement.
curiositechEncode identity, context, task, and protocol in four-branch hypertree prompts for DAG node execution.
niveshdandyanOrchestrate parallel agent swarms with per-agent handoffs and shared knowledge stores.
GuitarAlchemistSimulate multi-armed bandit algorithms with epsilon-greedy, UCB1, and Thompson sampling.
nicmartiFetch monster stats, generate encounters, and assign random hit points.
