Content
82%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 tight, highly executable API reference body with exemplary curl examples and a sensible auth-safe secret idiom, undermined only by its file layout: every promise of deeper detail (REFERENCE.md, TROUBLESHOOTING.md) resolves to a nonexistent file, so the skill's progressive-disclosure structure collapses at its first hop.
Suggestions
Create the referenced bundle files (REFERENCE.md with the endpoint list from src/triggers/api.ts, and TROUBLESHOOTING.md) or remove the links so no navigation path dead-ends.
Add a brief inline error-recovery note for failed curl calls (e.g., what a 401 vs 400 means and what to check) instead of relying solely on the missing TROUBLESHOOTING.md.
Include one or two additional endpoint examples inline (e.g., listing or deleting memories) so the quick start remains useful even while REFERENCE.md is being authored.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and assumes Claude's competence: "REST is agentmemory's primary surface. MCP is a bridge on top of it" and short sections (Quick start, Auth, Conventions, See also) with zero padding or explanations of concepts Claude already knows. Every token earns its place, matching the anchor "Lean and efficient; assumes Claude's competence". Not a 4 because there are no instances of over-explanation to trim. | 5 / 5 |
Actionability | Fully executable, copy-paste-ready curl commands cover the common cases: livez check, bearer-auth POST to `/agentmemory/remember` and `/agentmemory/smart-search`, plus a working secret-loading idiom (`agentmemory_secret="${AGENTMEMORY_SECRET:-$(cat "$HOME/.agentmemory/secret")}"`). Matches the anchor "Fully executable; copy-paste ready code or commands; specific examples cover the common cases". Not a 4 because there are no gaps in the shown commands; not below 5 since the endpoints shown are concrete and complete. | 5 / 5 |
Workflow Clarity | A coherent sequence is present (liveness check → load secret with an explicit guard `test -n "$agentmemory_secret" || exit 1` → save → recall), and status-code conventions ("Save returns `201`, reads return `200`, validation errors return `400`") act as implicit checkpoints. Matches "Clear sequence with most checkpoints present; minor validation gaps". Not a 5 because error recovery is delegated to a TROUBLESHOOTING.md link that does not exist, leaving a missing feedback loop; not a 3 because the sequence and the secret guard are explicit, and this is not a destructive/batch operation requiring the level-3 cap. | 4 / 5 |
Progressive Disclosure | Sections are well-signaled, but the only deep-dive pointers — "See [TROUBLESHOOTING.md](TROUBLESHOOTING.md)" and "The full endpoint list with methods lives in REFERENCE.md" — point to files that do not exist anywhere in the bundle (no references/, scripts/, or assets/ directories). Scoring against the actual bundle structure, navigation is broken: the reference surface the skill defers to is missing. Matches the anchor "Minimal structure; ... references are buried" in effect, since the promised reference layer is absent. Not a 3 because the problem is not unclear signaling but dangling references; not a 1 because the body itself is well organized with no nested chains. | 2 / 5 |
Total | 16 / 20 Passed |