skill-portabilitylisted

# Skill portability ## Goal Make a skill reusable by other users while preserving its behavior and keeping `SKILL.md` short. ## Input - the target skill and its nearby support files; - repository-level instructions such as `AGENTS.md` and `docs/`; - the current user's runtime/configuration when it can be inspected. ## Output A skill whose dependencies and configuration requirements are explicit, whose personal assumptions are removed or narrowly scoped, and whose setup is documented only as deeply as needed. ## Workflow 1. Read the target skill and its local docs/support files. 2. Find author-specific assumptions: repository owners, dotfiles, usernames, hosts, absolute paths, private services, config managers, MCP registrations, and undeclared commands. Use this skill's `scripts/audit.py <skill-dir> --literal <author-or-host>` helper for repeatable path/private-network/literal scanning; it reports categories and locations without echoing matched content. 3. Classify each assumption as a dependency, configuration detail, workflow rule, or optional owner optimization. 4. Rewrite dependencies as capabilities the skill actually consumes. 5. Replace personal paths/config with discovery, configurable values, or the current user's actual declarative source. 6. If configuration is missing and the user requested setup, inspect the user's environment and configure it when authorized and tooling permits. Do not impose the author's config system. 7. Put non-obvious install/config