Help us improve
Share bugs, ideas, or general feedback.
From codeharness
Integrates codeharness with BMAD methodology — reads sprint plans, maps stories to verification tasks, enforces harness requirements in all BMAD workflows. Triggers when working with BMAD artifacts, sprint plans, or story files.
npx claudepluginhub ivintik/private-claude-marketplace --plugin codeharnessHow this skill is triggered — by the user, by Claude, or both
Slash command
/codeharness:bmad-integrationThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
When a BMAD sprint plan exists at `_bmad-output/planning-artifacts/`:
Provides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Guides systematic root-cause debugging when tests fail, builds break, or unexpected errors occur. Provides a structured triage checklist to preserve evidence, localize, and fix issues instead of guessing.
Share bugs, ideas, or general feedback.
When a BMAD sprint plan exists at _bmad-output/planning-artifacts/:
codeharness patches these BMAD workflows (see templates/bmad-patches/):
Patches use idempotent markers (NFR19):
<!-- CODEHARNESS-PATCH-START:{name} -->
...
<!-- CODEHARNESS-PATCH-END:{name} -->
When BMAD is installed for OpenCode or another runtime without first-party BMAD tool integration:
npx bmad-method install --yes --directory . --modules bmm --tools none_bmad/ as the workflow artifact source of truthAGENTS.md, commands/, and skills/ files as the runnable agent-facing surface