Mak Sò

icon indicating a website link http://marcosomma.github.io/ icon indicating an email [email protected]

icon company @beehouse-network icon location Spain icon location Tech Polyglot and Innovation Enthusiast | Bridging the Gap Between Ideas and Reality | Embracing the Unconventional in Tech

Results 3 issues of Mak Sò

Add end-to-end (black-box) integration tests with deterministic LLM/memory stubs

Test coverage is strong but most integration or system tests use mocks for LLM and memory. Add real end-to-end tests that: - Run full example YAML workflows from CLI, exercising...

Implement real input readiness, safety, and priors scoring checks (remove placeholders)

Currently, scoring routines for input readiness, safety, and prior/historical success in PathScorer and related components are stubbed with fixed values and TODOs. For v1.0.0, replace these with: - Real input...

Expose boolean/deterministic scoring audit trail per path candidate

Current scoring is still float-based aggregates. For some compliance and auditable use cases, users want to know which criteria were discretely passed or failed by a given candidate path. Add:...