Content
86%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.
The body is a well-structured capability skill: concrete commands, clear gating on when to use the tool, and clean externalization of detail into two real reference files. Minor conciseness trims and the absence of an error-recovery feedback loop keep two dimensions at 4.
Suggestions
Tighten the 'What `trellis mem` is' and 'capability skill, not a workflow' framing to one or two sentences each to reclaim tokens without losing the key scoping message.
Add a brief note on what to do when `trellis mem` returns no matches or errors (e.g. CLI not installed / no indexed sessions), giving the recall flow a small feedback/checkpoint step.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Efficient and assumes Claude's competence (no padding about CLIs/JSONL), but a few sentences — the 'intentionally a capability skill, not a workflow' framing and the OpenCode caveat — could be trimmed without losing clarity. | 4 / 5 |
Actionability | Provides copy-paste-ready `trellis mem` commands with concrete flags (search, extract --phase, context --turns/--around, list --cwd, projects) covering the 80% common cases. | 5 / 5 |
Workflow Clarity | Clear sequence (when to reach → when not → what to do with output → how to call) with explicit gating ('If none of these apply, don't call mem'); appropriate for a read-only recall skill, with only minor validation gaps versus a full feedback-loop workflow. | 4 / 5 |
Progressive Disclosure | Clear overview body with two well-signaled one-level-deep references (`references/cli-quick-reference.md`, `references/triggering-patterns.md`), both real files, no nested references — content appropriately split for easy navigation. | 5 / 5 |
Total | 18 / 20 Passed |