What problem does it solve? Agent skills often accumulate scattered internal scripts and repeated mechanical instructions that waste tokens and behave inconsistently. This Skill analyzes one explicitly named target skill, moves its deterministic work into a single Python CLI, and rewires the target SKILL.md to call that one public interface. ## Core Features & Use Cases - Read-only analysis and classification: Inventories every tool in the target skill and tags each operation as CLI_REQUIRED, CLI_RECOMMENDED, OPTIONAL, KEEP_IN_SKILL, or NOT_JUSTIFIED, with an honest "no change" outcome allowed. - Unified CLI construction: Builds scripts/<skill-name>-cli.py with a hierarchical command tree, JSON output envelope, stable exit codes, --dry-run, self-check, and capabilities commands, consolidating legacy tools via WRAP, IMPORT, CONSOLIDATE, and other documented strategies. - Rewiring and behavioral validation: Updates the target SKILL.md and references to use the CLI, adds automated tests, and diffs behavior before and after to prove no workflow or semantic capability was lost. - Use Case: A maintainer notices a skill with five overlapping helper scripts and repeated shell pipelines. They explicitly invoke the Skill Toolsmith process on it, and receive one tested CLI plus a leaner SKILL.md with evidence-backed equivalence. ## Quick Start Run the Skill Toolsmith process on the target skill named project-archivist to consolidate its scripts into a single Python CLI.
