What problem does it solve? Manually testing a command-line accounting tool across every journey—directives, imports, reports, cloud sync—is slow and error-prone, and bugs often hide in edge cases like paths with spaces, partial batch writes, or JSON output inconsistencies. This Skill automates a disciplined CLI bug hunt that exercises the real executable, reproduces failures, and traces root causes. ## Core Features & Use Cases - Isolated test environments: Runs the bea CLI in subprocesses with dedicated config/cache directories and synthetic ledgers, never touching your real files, credentials, or shell state. - Full journey coverage: Sweeps discovery, directives, batch import, check/format, list/query, reports, automation/JSON output, cloud, and ask/upgrade journeys against documented behavior contracts. - Reproduce and trace: Repeats candidate bugs from pristine fixtures, distinguishes documented behavior from real defects, and traces root causes through the CLI source before filing deduplicated findings. - Use Case: Ask the agent to QA the reports journey on an installed bea binary; it builds fixed-date synthetic fixtures, compares output against independent BQL calculations, reproduces any discrepancy on the checkout, and reports findings by severity. ## Quick Start Use the qa-find-bugs-cli skill to hunt bugs in the bea CLI's reports and directives journeys using synthetic ledgers.
