SKILL · .zip
eval-runner
Score an action-plan against a 5-dimension rubric (LLM-judge, Bedrock Sonnet 4.5) and exercise its parallel-execution orchestrator against deterministic mocked sub-agents. Produces a scorecard JSON for archival and a markdown report for human consumption. Use when iterating on the action-plan skill, tuning the rubric, or qualifying a new book for action-plan.
$14.00
ONE-TIME PURCHASEv2
LATEST VERSION10
FILES§ FILES (V2)
The list below is the manifest — paths and sizes only. After purchase, install the whole skill with one command — mkdir -p ~/.claude/skills && curl -sL …/api/install/eval-runner | tar xz or, with the bkstr CLI, npx -y @clawbot678/bkstr install eval-runner — see Get Started below.
10 files in archive
ARCHIVE CONTENTS
- MDSKILL.md15.9 KB
- PYhelpers/__init__.py749 B
- PYhelpers/fetch_target.py6.1 KB
- PYhelpers/judge.py7.7 KB
- PYhelpers/mock_executor.py9.0 KB
- PYhelpers/pre_flight.py4.8 KB
- PYhelpers/render_report.py5.5 KB
- PYhelpers/score.py5.6 KB
- YAMLmocks/action_outcomes.yaml1.7 KB
- YAMLrubric.yaml6.1 KB