Content
100%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.
An exemplary short destructive-operation skill: it mandates disclosure and confirmation before a destructive command, provides the exact executable command with variants for each user intent, and closes with a report-back step. No section is wasted and no step is ambiguous.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Every line is operational — disclosure, confirmation gate, an executable command, flag variants, and reporting instructions — with zero padding and no explanation of concepts Claude already knows. "This permanently deletes remote memories" is a load-bearing warning, not filler. | 5 / 5 |
Actionability | The full command `python3 "${KIMI_PLUGIN_ROOT}/core/memory_cli.py" --harness "kimi" forget --remote --yes` is copy-paste ready, and the exact flag variations (`--include-project-memory`, dropping `--remote` for local-only) map each stated user intent to a concrete invocation. | 5 / 5 |
Workflow Clarity | Clear sequence — disclose what will be deleted, wait for explicit user confirmation, execute, report output — with a hard validation gate for a destructive operation ("Never pass `--yes` before the user has confirmed in this conversation"). The destructive-operation cap at 3 does not apply since confirmation is mandatory, and the single action is unambiguous per the simple-skill exception. | 5 / 5 |
Progressive Disclosure | At under 50 lines with no need for external references, the single heading and logically ordered paragraphs satisfy the rubric's simple-skill criterion; there are no bundle files and no buried references to navigate. | 5 / 5 |
Total | 20 / 20 Passed |