Thor Whalen
Community@thorwhalen · San Francisco, CA
Machine Learning and AI Systems
Agent Skills by Thor Whalen
Showing 9 vetted skills indexed across 3 GitHub repositories.
thorwhalenopenloops
Synthesizes ol command output into a briefing of open loops, obligations, and cross-repo blockers.
thorwhalenopenloops-needs-human
Files agent blockers as labeled GitHub issues carrying verifiable shell predicates.
thorwhalenek-dev-add-signal
Integrate custom signals, calibrators, and decision policies into the ek framework.
thorwhalenek-dev-add-metric
Wrap external libraries into standardized Protocol metrics for the ek framework.
thorwhalenek-dev-licensing
Evaluate library licenses against permissive, copyleft, and non-commercial criteria.
thorwhalenek-dev-architecture
Defines the architectural framework for Knowledge Evaluation systems with a two-layer data model.
thorwhalenek-dev-agents
Evaluate AI agent performance using cost per successful task and trajectory reliability metrics.
thorwhalenek-dev-ocr
Benchmark OCR engines against gold-standard references using CER, WER, and ANLS metrics.
thorwhalenfalaw
Generate and edit AI media via fal.ai with automatic model selection.
Frequently Asked Questions About Thor Whalen
FAQPage SchemaWhat tasks can I accomplish with thorwhalen's skills?▼
You can integrate custom signals, calibrators, and metrics into the ek knowledge-evaluation framework, benchmark OCR engines with CER/WER/ANLS scores, evaluate library licenses against permissive and copyleft criteria, measure agent cost per successful task, generate media via fal.ai, and track blockers across repositories with the `ol` command.
Who are these skills designed for?▼
They target machine learning engineers and developers building evaluation systems: teams benchmarking OCR output, assessing AI agent reliability and cost, auditing third-party library licenses, and developers who run many concurrent coding sessions and need a synthesized view of what needs their attention.
How do the openloops skills work in practice?▼
The openloops skill runs the read-only `ol` command and synthesizes its output to answer what needs attention across sessions and repos. The openloops-needs-human skill files blockers as `manual-task` GitHub issues carrying a shell predicate, so `ol owed` can re-check obligations later instead of losing them in chat.
What evaluation metrics do the ek-dev skills support?▼
The ek-dev-ocr skill benchmarks OCR engines using CER, WER, and ANLS against gold-standard references. The ek-dev-agents skill measures cost per successful task and trajectory reliability. The ek-dev-add-metric skill wraps external libraries into standardized Protocol metrics for the ek framework.
What are the prerequisites for using these skills?▼
The ek-dev skills require working within the ek knowledge-evaluation framework and its two-layer data model architecture. The openloops skills require the `ol` command and GitHub repositories where manual-task issues can be filed. The falaw skill requires access to fal.ai for media generation.