Content
67%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 lean, largely executable reference for a simple CLI skill: concrete commands, a full options table, and covering examples. Its weaknesses are redundant duplication between Common Patterns and Examples, and a muddled final workflow whose last step does not match the state it receives (an agent_id, not a partial name).
Suggestions
Merge the Examples section into Common Patterns (or keep one) to eliminate the repeated commands and cut the token cost of the skill.
Fix the topic workflow's final step to act on the agent_id actually obtained, e.g., a concrete command like "letta agents list --name <partial-name>" or a documented agents-get endpoint instead of the vague "use the agent_id directly in the Letta API".
Promote "Finding Agents by Topic" to its own top-level section (or nest it under CLI Usage) rather than burying it beneath "## Output", since it is a workflow, not an output-field description.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The Common Patterns and Examples sections repeat the same commands verbatim ("letta agents list", "--shared", "--tags \"origin:letta-code\"\" each appear twice), and filler like "This is useful when the user is looking for agents they've worked with in Letta Code CLI sessions" adds little. Mostly efficient but clearly could be tightened by merging the duplicated sections, matching the anchor of 3 rather than the minor-trim anchor of 4. | 3 / 5 |
Actionability | Commands are copy-paste ready with a complete options table and six concrete examples covering common cases. The gap is in the topic workflow, where step 3 suggests "letta agents list --query \"partial-name\"" even though the user holds an agent_id, and "use the agent_id directly in the Letta API" gives no executable command — minor gaps that keep it below the fully-executable anchor of 5. | 4 / 5 |
Workflow Clarity | The core single action (letta agents list) is unambiguous, and the topic workflow is a numbered 3-step sequence. Step 3's guidance mismatches the workflow state (querying by partial name when an agent_id is in hand), a minor gap; operations are read-only so the destructive/batch validation cap does not apply. | 4 / 5 |
Progressive Disclosure | No bundle files exist and the skill is organized into clear sections with nothing nested more than one level deep, which is appropriate for a skill of this size. Minor organizational gaps remain: "Finding Agents by Topic" is oddly nested as a subsection under "## Output", and the Examples section largely duplicates Common Patterns instead of pointing to it. | 4 / 5 |
Total | 15 / 20 Passed |