Use this skill for ALL build, test, and CI operations. When the user asks to run tests, build, check lint, run type checking, or analyze build output, route through blq MCP tools instead of running commands through Bash. Also use when analyzing errors, filtering output, or checking CI status.
blq captures and queries build/test output. Use blq MCP tools instead of running build commands through Bash directly.
output() to filter after the factDo NOT run build/test commands through Bash. Use blq MCP tools instead:
| Instead of... | Use blq |
|---|---|
pytest tests/ | mcp__blq_mcp__run(command="test") |
make build | mcp__blq_mcp__run(command="build") |
ruff check src/ |
mcp__blq_mcp__run(command="lint") |
mypy src/ | mcp__blq_mcp__run(command="typecheck") |
| Piping output through tail/grep | Run first, then use mcp__blq_mcp__output(tail=20) or output(grep="FAIL") |
| Ad-hoc shell commands | mcp__blq_mcp__exec(command="any shell command") |
mcp__blq_mcp__commands() # list available commands
mcp__blq_mcp__run(command="test") # run a registered command
mcp__blq_mcp__run(command="test", lines="+20-") # run and get last 20 lines inline
mcp__blq_mcp__status() # current build/test status
mcp__blq_mcp__errors() # view errors from recent runs
mcp__blq_mcp__output(tail=20) # last 20 lines of output
mcp__blq_mcp__output(grep="FAIL") # search output
mcp__blq_mcp__info(ref="latest") # detailed run info
mcp__blq_mcp__events() # all events (errors, warnings, info)
mcp__blq_mcp__inspect(ref="latest") # inspect a specific run
mcp__blq_mcp__history() # run history
mcp__blq_mcp__diff() # compare runs
output()commands() first — see what's registered before runningrun() not Bash for registered commands — run captures and indexes outputlines parameter on run() for quick inline output (e.g., lines="+20-" for last 20 lines)