Content
92%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 tightly engineered operational protocol: concrete metric definitions, file paths, routing rules, and a done-when checklist leave little ambiguity about what to execute, and all depth is externalized to clearly annotated one-level references. Its only weakness is repetition — the ECHO/scope boundary is disclaimed in four places — which costs a little token efficiency without harming clarity.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body assumes Claude's competence — no concept explanations, only operational protocol — and every section carries instructions ("Median, not mean", "boosted numbers never enter organic trend lines"). However, the ECHO/scope boundary is restated at least four times (intro, Scope guard, step 7, Next Best Skill), which is trimmable repetition, matching 'Efficient; minor instances of over-explanation that could be trimmed' rather than the lean anchor 5, and it is well above the verbosity of anchor 3 since no genuinely unnecessary explanation exists. | 4 / 5 |
Actionability | Guidance is fully executable for an instruction-only skill: exact formulas ("ERR = engagements ÷ reach, ERI = engagements ÷ impressions, ER-by-follower = engagements ÷ followers"), a concrete save path template (`memory/social/social-measurement-loop/YYYY-MM-DD-<period>-readout.md`), named connector scripts (`discourse.py`, `bluesky.py`, `experiment.py proportion`), explicit event-routing mechanics (`operation: propose` to `registry-events.py`), and labeled status values (`binding_status: incomplete`). Per the rubric's code-vs-instruction note, absence of code is not penalized when guidance is this actionable; this matches 'Fully executable; copy-paste ready... specific examples cover the common cases'. | 5 / 5 |
Workflow Clarity | The eight numbered Instructions form a clear sequence (bind period → dictionary → rollup → deltas → EMV → community-health → routing → write-back) with explicit validation: the "Done when" checklist ("every included post has a matching human publication receipt... every reported rate names and preserves its denominator... EMV appears in no score") and failure handling ("A missing receipt or incomplete window yields a provisional hypothesis/open loop, not a terminal content decision"). This matches 'Clear sequence with explicit validation steps; feedback loops...; checklists'; the operation is read/analyze with confirmation-gated writes, so no destructive-operation cap applies. | 5 / 5 |
Progressive Disclosure | No bundle files ship in this skill's directory (references/, scripts/, assets/ are absent), and the body is a well-organized overview: Quick Start examples, a Skill Contract, and a Reference Materials section where every link is one level deep and annotated with its purpose (e.g., "echo-benchmark.md — the O sub-items this skill feeds and the ECHO-O1 denominator-integrity veto"). This matches 'Clear overview with well-signaled one-level-deep references; content appropriately split; easy navigation' — no detail bulk is inlined that belongs in a separate file. | 5 / 5 |
Total | 19 / 20 Passed |