Distill best practices from agent-runtimes M1-M3 memory files
12 additions/updates across 5 best-practice files:
- docker-uid-matching: userdel simplification, SSH agent socket UID match
- debugging: GIT_SSH_COMMAND scope limitation
- test-driven-development: subprocess mock gotcha, routing callables,
Pydantic v2 field_validator defaults, sys.exit at module level
- spec-driven-development: multi-agent orchestration practices (commit WIP,
self-verify, import conventions, assembly budget)
- validation: test pre-commit hooks after adding dependencies
Source: agent-runtimes/memory/ (decisions, gotchas-docker, gotchas-python,
process-lessons, m1/m2/m3 reflections)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
1b5e73dc54