Repository navigation
v0.13.0
Two features: the LLM-judge tools are now on the CLI (qulib score-decisions with a --min-quality CI gate, qulib score-bug-report), and qulib validate (+ MCP qulib_validate_spec) grades a deployed app against a supplied spec (PRD / requirements) — does it do what it was supposed to do? Honest insufficient-evidence without a key; untrusted spec/observed text is prompt-injection hardened. Publishes @qulib/core then @qulib/mcp at 0.13.0 via OIDC trusted publishing.