Content
86%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 well-structured, token-efficient review skill with concrete commands and a strong output template. The gaps are operational: script invocations lack arguments, and the workflow has no explicit validation or data-availability checkpoint.
Suggestions
Show the expected invocation for each analysis script, including arguments (e.g., input plan path or feature list), so the commands are copy-paste ready rather than bare paths.
Add a validation checkpoint in the Workflow, e.g. a minimum cohort sample size before judging a retention curve 'decaying', and an instruction for what to do when PMF/retention data is unavailable.
Map each script's output to the question it answers (e.g., 'pmf_scorer.py → Question 3; rice_prioritizer.py → Question 4') so the analyses visibly feed the six questions.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and punchy — questions are 2-3 lines with calibration examples ("Not 'improve onboarding.'"), and it never explains concepts Claude already knows (no JTBD or RICE primer). Every section earns its tokens; matches anchor 5 rather than 4, which requires trims. | 5 / 5 |
Actionability | Includes runnable commands ("python ../../../c-level-advisor/skills/cpo-advisor/scripts/pmf_scorer.py") and a concrete fill-in-the-blank output template with a verdict enum. Not 5: the script commands take no arguments, and nothing ties each script's output to the specific question it informs. | 4 / 5 |
Workflow Clarity | The Workflow section sequences run-analyses → answer-questions → apply-verdict with the output template acting as a checklist (cohort sample size, thresholds, verdict). Not 5: there is no explicit validation or recovery step (e.g., what to do when retention data or a script is unavailable), leaving checkpoints implicit. | 4 / 5 |
Progressive Disclosure | No bundle files exist, and the self-contained ~105-line body is well-sectioned with nothing that belongs in a separate file; the Related/Routing links are clearly signaled and one level deep. Matches anchor 5's clear overview and easy navigation. | 5 / 5 |
Total | 18 / 20 Passed |