Content
52%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 well-sequenced and unusually actionable for a thinking-framework skill, with concrete question sets, formats, and a decision-record template. Its two real weaknesses are verbosity — long stretches re-explain standard concepts Claude already knows — and a progressive-disclosure structure whose reference links are broken or mis-rooted, leaving the inlined bulk as the only usable content.
Suggestions
Fix the reference structure: create the four modules/ files (or relink to existing files) and correct 'examples.md'/'reference.md' to references/examples.md and references/reference.md so the Quick Access and deep-dive links resolve.
Trim concept explanations Claude already knows — the bias definitions, Five Whys walkthrough, and generic assumption categories — down to the operational question lists and formats, and remove the invented percentage statistics.
Add an explicit recovery loop to the workflow, e.g. 'If any Verification checklist item fails, return to the corresponding phase and re-run it before finalizing the decision record.'
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The ~330-line body spends much of its length explaining concepts Claude already knows — definitions of anchoring, confirmation bias, and sunk cost ('Over-reliance on first information encountered', 'Continuing due to past investment'), a walkthrough of the Five Whys, and generic assumption categories — plus padded invented statistics ('prevents 40-60% of mid-project pivots', 'reduces solution complexity by 30-50%'). This is noticeably verbose with several padded sections, matching the 2 anchor rather than the mostly-efficient 3 anchor. | 2 / 5 |
Actionability | For an instruction-only skill the guidance is concrete: per-phase question lists ('What if this assumption turns out to be wrong?'), a five-field assumption documentation format, a weighted scoring method with explicit steps, and a full Strategic Decision Record template. Minor gaps keep it below 5 — the scoring method has no worked example with numbers, and some sections (e.g. the creativity techniques) stay at the level of named prompts rather than executable procedure. | 4 / 5 |
Workflow Clarity | The five phases are explicitly sequenced with per-phase purposes, reinforced by a Quick Decision Matrix mapping task types to phases, time-allocation guidelines, and a closing Verification checklist ('At least two alternative approaches documented before selecting one'). It falls short of the 5 anchor because there is no error-recovery loop — nothing says what to do when a verification item fails or when to revisit an earlier phase. | 4 / 5 |
Progressive Disclosure | The three-tier reading structure (30 seconds / 5 minutes / 10+ minutes) is a good idea, but the actual bundle does not support it: the four module links ('modules/assumption-matrix.md', 'modules/first-principles.md', etc.) point to files that do not exist, and 'examples.md' and 'reference.md' are linked at the skill root while the real files live under references/. Meanwhile the full detail that supposedly lives in the modules is inlined in the 5-minute section. This matches the 2 anchor — content that belongs in separate files is inlined and references are effectively broken — rather than the partially-organized 3 anchor. | 2 / 5 |
Total | 12 / 20 Passed |