Content
88%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 highly actionable, well-sequenced conversational skill: every step has copy-paste-ready payloads or commands, all three modes are cleanly dispatched, and destructive operations are gated behind confirmations and a dry-run loop. The only notable weaknesses are repeated language-handling rules and a long single file where some verbatim copy could be split out.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense with copy-paste ask_user JSON payloads, exact shell commands, and templates, and never explains concepts Claude already knows. However, the language-handling rules are stated four times (Dispatch intro, A.1, B.1, C.2) and the full bilingual celebration copy is inlined, so it could be tightened. | 4 / 5 |
Actionability | Every step ships executable material: exact card JSON in both languages, exact commands like `ruby "SKILL_DIR/scripts/install_builtin_skills.rb"` and `test -d ~/.openclaw && echo yes || echo no`, and complete SOUL.md/USER.md templates. Nothing is pseudocode or hand-wavy. | 5 / 5 |
Workflow Clarity | All three modes are sequenced in numbered subsections with an explicit dispatch table and jump instructions. Risky operations carry validation checkpoints: import runs dry-run then confirm then --yes, memory deletion requires explicit confirmation and uses the trash tool rather than raw rm, and B.7 confirm-and-write loops back to B.4 on tweak. | 5 / 5 |
Progressive Disclosure | Executable logic is properly pushed into two scripts referenced as SKILL_DIR/scripts/... (both verified to exist), one level deep and clearly signaled. The main file is ~580 lines with sizable inlined verbatim copy (bilingual cards and celebration scripts) that could arguably live in reference files. | 4 / 5 |
Total | 18 / 20 Passed |