Content
90%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 lean, executable API-surface skill: copy-paste curl examples for the common operations, concrete auth and status-code conventions, and clean deferral of the bulk endpoint list. The two real weaknesses are a missing REFERENCE.md in the bundle (the primary deferred payload cannot actually be read) and the absence of any error-recovery guidance around the quick-start sequence.
Suggestions
Ship REFERENCE.md in the bundle (e.g., under references/) — the body defers the full endpoint list to it, but no such file exists, leaving the skill's core payload unreachable.
Add a two-line error-recovery note (e.g., 401 → AGENTMEMORY_SECRET not set / passed; connection refused → check --port or --instance) to close the validation loop in the Quick start.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The ~40-line body is lean: five short sections, no explanation of REST/HTTP basics, and the one positioning sentence ("REST is agentmemory's primary surface. MCP is a bridge on top of it.") is project-specific knowledge Claude cannot infer. Every section earns its place; no over-explanation was found that would justify anchor 4. | 5 / 5 |
Actionability | "curl -fsS http://localhost:3111/agentmemory/livez", the POST /remember and POST /smart-search examples with full JSON bodies, the concrete auth header ("Authorization: Bearer $AGENTMEMORY_SECRET"), and the 201/200/400 status conventions are copy-paste ready and cover the common cases (liveness, save, recall). Matches anchor 5. | 5 / 5 |
Workflow Clarity | The Quick start sequences a liveness check before save/recall — an explicit validation checkpoint — and status-code conventions give feedback on outcomes. However, there is no error-recovery loop (e.g., 401 → AGENTMEMORY_SECRET not set; connection refused → wrong --port), which keeps it at anchor 4 ("most checkpoints present; minor validation gaps") rather than 5. It is a reference-style skill with no destructive or batch operations, so the workflow-clarity cap at 3 does not apply. | 4 / 5 |
Progressive Disclosure | The body is a clean overview that defers bulk content appropriately ("The full endpoint list with methods lives in REFERENCE.md") with clearly signaled, one-level-deep references plus sibling-skill "See also" pointers. However, no bundle files exist (no references/, scripts/, or assets/ directories) and REFERENCE.md is not present anywhere in the bundle, so the skill's core payload — the full endpoint list — is unreachable. The structure itself is anchor-5 quality; the dangling primary reference is more than a minor organization gap but the in-file organization and signaling are solid, placing it between anchors 4 and 5, at 4. | 4 / 5 |
Total | 18 / 20 Passed |