John Morrissey
Community@tachyon-beep · Canberra, ACT
I am neurodiverse, projects aren't abandoned. I've just become distracted by something else for now.
Agent Skills by John Morrissey
Showing 55 vetted skills indexed across 2 GitHub repositories.
tachyon-beepvalidating-architecture-analysis
Validate architecture artifacts against contracts and cross-document consistency standards.
tachyon-beepanalyzing-unknown-codebases
Analyze unfamiliar codebases and output standardized subsystem catalog entries.
tachyon-beepdocumenting-system-architecture
Synthesize subsystem catalogs and diagrams into structured architecture reports.
tachyon-beepgenerating-architecture-diagrams
Translate subsystem catalogs into C4 architecture diagrams with Mermaid or PlantUML.
tachyon-beepidentifying-technical-debt
Catalog prioritized technical debt items with explicit scope and delivery dates.
tachyon-beepprioritizing-improvements
Enforce risk-based prioritization with a security-first Phase 1 roadmap.
tachyon-beepassessing-architecture-quality
Assess codebase architecture quality with evidence and severity ratings.
tachyon-beeppolicy-gradient-methods
Apply REINFORCE, PPO, and TRPO to optimize policies in continuous-action tasks.
tachyon-beeprl-debugging
Diagnose reinforcement learning training failures with structured diagnosis trees.
tachyon-beepmodel-based-rl
Study model-based RL methods with world models and planning.
tachyon-beepreward-shaping-engineering
Design reward functions that preserve optimal policies in reinforcement learning.
tachyon-beepmulti-agent-rl
Implement QMIX and MADDPG algorithms for multi-agent reinforcement learning with PyTorch.
tachyon-beepactor-critic-methods
Guide selection, configuration, and debugging of actor-critic methods for continuous-control reinforcement learning.
tachyon-beeprl-evaluation
Evaluate RL agents across multiple seeds with mean, std, and confidence intervals.
tachyon-beepoffline-rl
Implement CQL, IQL, and BCQ for offline reinforcement learning from fixed datasets.
tachyon-beeprl-foundations
Explain MDPs, value functions, Bellman equations, and policy optimization.
tachyon-beepexploration-strategies
Implement and compare exploration strategies for deep RL agents.
tachyon-beepvalue-based-methods
Guide implementation of DQN variants for discrete action spaces.
tachyon-beeprl-environments
Design and validate custom Gym/Gymnasium environments with spaces, wrappers, and vectorization.
tachyon-beepusing-ai-engineering
Route AI/ML engineering tasks to the correct Yzmir pack by problem type.
tachyon-beepusing-ml-production
Route ML production concerns to relevant deployment, optimization, MLOps, and observability sub-skills.
tachyon-beepusing-llm-specialist
Route LLM tasks to specialized skills via a decision tree.
tachyon-beepusing-security-architect
Route security tasks to threat-modeling, secure-by-design, and security-controls reference sheets.
tachyon-beepusing-python-engineering
Route Python problems to the appropriate specialist skill by symptom classification.