I took a look at how you've structured the product-manager-toolkit skill—the way you've organized the mental models and frameworks for different PM scenarios is pretty thoughtful. What made you decide to go heavy on the decision-making frameworks rather than process templates?
Links:
The TL;DR
You're at 80/100, solid B territory. This is graded against Anthropic's best practices for token efficiency, ease of use, and practical utility. Your strongest area is Utility (17/20)—the RICE prioritizer and PRD tools actually solve real PM problems. Weakest area is Writing Style (7/10)—some verbose sections and inconsistent voice that's holding you back from the high 80s.
What's Working Well
- Practical Python tools - The rice_prioritizer.py and prd_templates.md actually generate working output; not just theory
- Clear trigger phrases - "RICE prioritization", "PRD", "customer interview" make it easy to know when to use this
- Good framework scaffolding - RICE formula, MoSCoW, Opportunity Solution Tree are well-explained in context
- Modular structure - References and scripts are cleanly separated from the main SKILL.md
The Big One: Missing TOC + Verbose SKILL.md
Your SKILL.md is 352 lines without a table of contents—that's a hard miss on Progressive Disclosure Architecture (PDA). More importantly, you're burying the lead. Half that content (RICE formula, MoSCoW definitions, customer interview guides, best practices) belongs in references/frameworks.md.
The fix: Add a TOC to SKILL.md, then move the detailed frameworks to a new reference file. Keep SKILL.md as the cockpit (overview + quick start), let references hold the manuals. This'll tighten your token usage and bump PDA from 22 to 25+.
Other Things Worth Fixing
-
Second-person voice creeping in (lines 27, 104, 335) - You're saying "your PM tool" and "your environment" instead of imperative form. Stay consistent: "Integrates with..." not "Your toolkit integrates..."
-
No input/output examples - Show what users will actually see. Add a concrete example: Input CSV with 3 features, output showing RICE scores and priority ranking. This is quick but high-impact for Utility.
-
Incomplete feedback loops - Your workflows stop at "Generate Roadmap" but don't show validation. Add a "Check Results" step: compare against strategic goals, sensitivity analysis, stakeholder review → adjust → regenerate.
-
Redundant framework sections - RICE is explained both in Core Workflows (lines 39-57) AND in a standalone section (lines 176-194). Pick one location, eliminate the duplicate.
Quick Wins
- Add TOC to SKILL.md → +2 points (PDA)
- Move frameworks to references/frameworks.md → +3 points (PDA)
- Add input/output examples → +1 point (Utility)
- Fix second-person voice → +1 point (Writing Style)
- Consolidate duplicate framework definitions → +1 point (PDA)
That's realistically +8 points to hit 88/100 with minimal refactoring.
Checkout your skill here: SkillzWave.ai | SpillWave We have an agentic skill installer that install skills in 14+ coding agent platforms. Check out this guide on how to improve your agentic skills.
I took a look at how you've structured the product-manager-toolkit skill—the way you've organized the mental models and frameworks for different PM scenarios is pretty thoughtful. What made you decide to go heavy on the decision-making frameworks rather than process templates?
Links:
The TL;DR
You're at 80/100, solid B territory. This is graded against Anthropic's best practices for token efficiency, ease of use, and practical utility. Your strongest area is Utility (17/20)—the RICE prioritizer and PRD tools actually solve real PM problems. Weakest area is Writing Style (7/10)—some verbose sections and inconsistent voice that's holding you back from the high 80s.
What's Working Well
The Big One: Missing TOC + Verbose SKILL.md
Your SKILL.md is 352 lines without a table of contents—that's a hard miss on Progressive Disclosure Architecture (PDA). More importantly, you're burying the lead. Half that content (RICE formula, MoSCoW definitions, customer interview guides, best practices) belongs in
references/frameworks.md.The fix: Add a TOC to SKILL.md, then move the detailed frameworks to a new reference file. Keep SKILL.md as the cockpit (overview + quick start), let references hold the manuals. This'll tighten your token usage and bump PDA from 22 to 25+.
Other Things Worth Fixing
Second-person voice creeping in (lines 27, 104, 335) - You're saying "your PM tool" and "your environment" instead of imperative form. Stay consistent: "Integrates with..." not "Your toolkit integrates..."
No input/output examples - Show what users will actually see. Add a concrete example: Input CSV with 3 features, output showing RICE scores and priority ranking. This is quick but high-impact for Utility.
Incomplete feedback loops - Your workflows stop at "Generate Roadmap" but don't show validation. Add a "Check Results" step: compare against strategic goals, sensitivity analysis, stakeholder review → adjust → regenerate.
Redundant framework sections - RICE is explained both in Core Workflows (lines 39-57) AND in a standalone section (lines 176-194). Pick one location, eliminate the duplicate.
Quick Wins
That's realistically +8 points to hit 88/100 with minimal refactoring.
Checkout your skill here: SkillzWave.ai | SpillWave We have an agentic skill installer that install skills in 14+ coding agent platforms. Check out this guide on how to improve your agentic skills.