Content
81%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 strong, well-structured body: the two-mode workflow is explicit with verification checkpoints and troubleshooting feedback loops, and the examples are genuinely instructive rather than filler. The main weakness is redundancy — the Core Mechanics section and some examples repeat both the workflow steps and material delegated to the reference files, which costs tokens without adding guidance.
Suggestions
Cut or merge the Core Mechanics Reference section: its bullets restate the numbered Workflow steps for both modes, and its topics are already covered in references/character-cringe-craft.md and references/provocateur-standup-craft.md per the References table — keeping one canonical statement (the workflows) would remove the largest duplication.
Deduplicate examples: the 'honorary member of every minority' scene appears three times and the wedding-registry bit twice; keep each in the single section where it teaches most (Anti-Patterns) and drop the repeats from Core Mechanics/Examples.
Trim the Anti-Patterns 'Consequence of skipping this rule' lines, which restate the WHY paragraph immediately above them in each entry.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly efficient — it teaches craft mechanics Claude does not already have — but there is real duplication: the "honorary member of every minority" example appears three times (Core Mechanics example, Anti-Patterns BAD and GOOD), the wedding-registry example appears twice (Examples and Anti-Patterns), and the Core Mechanics Reference largely restates the numbered Workflow steps for both modes. This fits the anchor 'mostly efficient but includes some unnecessary explanation or could be tightened' rather than the level-4 anchor, whose over-explanation would be only minor. | 3 / 5 |
Actionability | For an instruction-only skill the guidance is fully executable: numbered per-mode workflows ("Build the authority figure around one specific, nameable delusion", "Write the opening line as one unhedged sentence... strip qualifiers"), six before/after text examples showing exact lines, and symptom→fix Troubleshooting entries. Per the rubric's code-vs-instruction note, absence of code is not penalized when guidance is this concrete and example-covered — matching the top anchor. | 5 / 5 |
Workflow Clarity | Both mode workflows are clearly sequenced (prerequisites → numbered steps → Verification), the Verification section is an explicit five-item checklist with a re-read pass ("Read it back once more start to finish: does the discomfort stay unresolved..."), and Troubleshooting provides feedback loops for error recovery. This matches the anchor with explicit validation steps, feedback loops, and checklists; level 4 would require missing checkpoints, which are present. | 5 / 5 |
Progressive Disclosure | Structure is good: two real, one-level-deep reference files (character-cringe-craft.md, provocateur-standup-craft.md) linked inline at the exact workflow steps that need them, plus a References table with per-reference "When to Use" guidance including when to skip them. It falls short of the 5 anchor because the ~60-line Core Mechanics Reference and several of the six worked examples overlap content the reference files already cover (the table itself lists "delusion design, onlooker roster, escalation ladder, comeuppance sizing" as reference-file material), so a modest amount of content that belongs in the bundle is inlined. | 4 / 5 |
Total | 17 / 20 Passed |