b-amir
- 4 skills
- 0 followers
- 22 hours ago last updated
- ▌ Repay Techdebt · b-amir bundleLearn this codebase from its live files through evidence-qualified lessons and a browser workbook for unfamiliar or agent-authored code. Prefer optional tools silently when already available. Otherwise use bundled scripts with the same UX (no tool jargon in user chat, no confidence inflation). Ask only before install/config changes. Hard overclaim rule: unsupported → shrink scope or refuse. Never offer "continue weaker?". Scripts and the agent take turns. Follow nextAsks on script output.
- ▌ Product Playbook · b-amir bundleCreate, reconcile, audit, or agent-check evidence-backed manual testing playbooks from any product codebase (tests, contracts, Swagger, docs, source, live interfaces). Use when the user asks for a product playbook, manual QA playbook, UAT script, tester-facing scenarios, playbook reconciliation, playbook audit, or drift check against an existing playbook. Do not use for writing unit tests, designing APIs, or generic QA strategy without a repository. Never invent interface labels or behavior. Publish only tester-facing Markdown plus one portable state JSON; optional PDF/HTML and sibling findings only when asked.
- ▌ Product Playbook Audit · b-amirAudit an existing product playbook against current evidence without editing Markdown or state. Use when the user asks for a playbook audit, drift check, coverage gap review, or read-only reconciliation report. Prefer the main product-playbook skill for create or edit runs.
- ▌ Product Playbook Export · b-amirExport an existing validated product playbook to a single PDF or HTML file. Use only when the user explicitly asks for playbook.pdf or playbook.html. Never invent playbook content; never run during normal create or reconcile.