Content
65%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.
This is a well-organized, highly actionable CLI reference with executable examples throughout and genuinely useful non-obvious details (subscriber hashing, rate limits). Its weaknesses are the inlined bulk command catalog that belongs in a reference file and the absence of safety/validation guidance around destructive and batch operations, plus one inconsistent command name in the patterns section.
Suggestions
Move the full command tables (Root through Other Resources, ~110 lines) into a references/COMMANDS.md file, keeping only the most common commands and patterns inline.
Add validation guidance for destructive/batch operations, e.g. verify a campaign's send-checklist before `campaigns send`, and confirm list/campaign IDs before delete commands.
Fix the command-name inconsistency: `lists create-members` in Common Agent Patterns should match the documented `lists create-lists-id-members`.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense reference material — terse command tables, envelope shapes, and non-obvious notes (MD5 subscriber hash, datacenter suffix) with no padding about concepts Claude already knows. Minor instances of trimmable content remain (the ~13-line REPL transcript and the Root table), so it fits 'efficient; minor over-explanation' rather than 'every token earns its place'. | 4 / 5 |
Actionability | Commands are copy-paste ready with real JSON payloads, jq pipelines, and a working MD5 hash one-liner. One concrete gap: the pattern `lists create-members <list_id>` contradicts the command table's `lists create-lists-id-members <LIST_ID>`, so that example would not execute as written — 'mostly executable with minor gaps'. | 4 / 5 |
Workflow Clarity | Multi-step agent patterns exist (create campaign → send-checklist → filter unmet items), but destructive and batch operations (`lists delete`, `campaigns send`, the `batches` group) have no validation or confirmation guidance. The rubric explicitly caps workflow clarity at 3 for batch/destructive skills missing feedback loops. | 3 / 5 |
Progressive Disclosure | Section headers and tables give the body decent structure, but ~110 lines of full command catalog are inlined in SKILL.md with no bundle files at all — content that clearly belongs in a one-level-deep reference file, matching 'content that should be separate is inline'. | 3 / 5 |
Total | 14 / 20 Passed |