Content
81%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A well-engineered workflow skill: concrete paths, templates, and verification commands, with a clearly sequenced seven-step process that includes explicit validation and failure-recovery checkpoints for batch and destructive operations. The main improvement opportunities are literal clone commands instead of descriptive patterns and slight trimming of the duplicated manifest example.
Suggestions
Replace the descriptive clone pattern in Step 5 with literal commands (e.g., `git clone --depth 1 <repoUrl> <tmpdir>` followed by the checkout and move), making the core operation copy-paste executable.
Trim the manifest example to one full dependency entry plus a note that additional entries share the same shape, saving tokens without losing schema information.
Consider moving the verbatim explorer prompt to a references/ file and pointing to it, keeping SKILL.md as a leaner overview.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is directive and dense throughout — 'Do not use a helper script for dependency detection, ref validation, cloning, status, or cleanup', 'Do not run dependency install/build/test scripts from cloned repositories' — with no explanations of concepts Claude already knows. Minor trimming is possible (the manifest example repeats two near-identical dependency entries where one plus a comment would do; the explorer prompt is long), which keeps it just below the lean score-5 anchor. | 4 / 5 |
Actionability | Guidance is highly concrete: exact gitignore block, path scheme `.slim/clonedeps/repos/<safe-repo-name>/` with a worked example (`opencode-ai__opencode`), verification commands (`git ls-remote <repoUrl> <ref>`, `git remote get-url origin`), a full JSON manifest example, and a copy-paste AGENTS.md template. It falls short of the fully copy-paste-ready score-5 anchor because the core clone step gives a pattern ('Clone without submodules', 'Prefer shallow fetch/clone') rather than literal commands like `git clone --depth 1 <url> <tmpdir>`. | 4 / 5 |
Workflow Clarity | Seven clearly sequenced steps with explicit validation checkpoints and recovery loops for batch/destructive operations: verify refs with `git ls-remote` before cloning, 'verify that `git remote get-url origin` matches the approved repo URL. If it does not match, stop and ask whether to clean/reclone', ask for confirmation before network cloning, 'Remove failed temporary clones', and 'If a clone fails after earlier clones succeeded, still write state for the successful clones'. This matches the score-5 anchor (explicit validation, error-recovery feedback, confirmation gates) and satisfies the destructive/batch validation requirement. | 5 / 5 |
Progressive Disclosure | The skill is a single self-contained file with no bundle files, and the body is well organized into numbered workflow steps plus a cleanup section, so nothing is buried and navigation is easy. The inline ~35-line explorer prompt, JSON manifest example, and AGENTS.md template could arguably move to reference files, which is the 'minor organization gap' of the score-4 anchor rather than the score-5 'clear overview pointing to one-level-deep references' structure. | 4 / 5 |
Total | 17 / 20 Passed |