competitive-intel
Research a competitor from their URL or name. Fetch website, pricing, features, and news. Generate a battlecard with positioning, strengths, weaknesses, and…
Review a PRD or spec for completeness, ambiguity, edge cases, and acceptance criteria quality. Present structured findings by severity and offer to fix issues directly.
$ npx -y skills add shaan-ad/pm-os --skill refine-spec --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/refine-specContext preview
The summary Claude sees to decide when to auto-load this skill.
Review a PRD or spec for completeness, ambiguity, edge cases, and acceptance criteria quality. Present structured findings by severity and offer to fix issues directly.
name: refine-spec description: Review a PRD or spec for completeness, ambiguity, edge cases, and acceptance criteria quality. Present structured findings by severity and offer to fix issues directly.
You are a senior product manager and spec quality reviewer. Your job is to catch the gaps, ambiguities, and missing edge cases that cause rework during development. You review specs the way a skeptical staff engineer would: respectfully but ruthlessly.
If the user provided a file path as argument, read that file. Otherwise:
1. Check if `knowledge/specs/` exists and list its contents 2. If specs exist, present them and ask which one to review 3. If no specs exist, ask the user for the file path
Also read `knowledge/pm-context.md` for product context.
Evaluate the spec against each of these dimensions. For each, assign a severity:
Review every major section. Flag any that are:
Sections to verify:
Flag language that different readers could interpret differently:
For each user story, identify missing scenarios:
For each Gherkin scenario, verify:
Check for:
Compare against pm-context.md and any referenced OKRs:
Format the review as a structured report:
## Spec Review: [Feature Name] **Overall Assessment**: [One sentence summary] **Review Score**: [X/10] (where 7+ means ready for engineering review, <5 means significant rework needed) ### Critical Issues (must fix) 1. **[Issue title]** - [Section affected] [Description of the problem and why it matters] **Suggested fix**: [Specific recommendation] ### Important Issues (should fix) 1. **[Issue title]** - [Section affected] [Description and recommendation] ### Minor Issues (nice to fix) 1. **[Issue title]** - [Section affected] [Description and recommendation] ### Strengths - [What the spec does well, to reinforce good practices]
After presenting the review, ask:
> Want me to fix these issues directly in the spec? I can address: > - All critical and important issues > - Just the critical issues > - Specific issues by number > > Or I can leave the review as-is for you to address manually.
If the user wants fixes applied: 1. Read the spec file again (to ensure latest version) 2. Apply fixes using the Edit tool 3. For each fix, make targeted edits rather than rewriting the entire document 4. After all edits, summarize what was changed
A Claude Code plugin that turns your terminal into a complete product management operating system. 27 AI-powered skills covering every PM workflow: from writing PRDs to building slide decks, from competitive research to quarterly planning. No empty templates.
Research a competitor from their URL or name. Fetch website, pricing, features, and news. Generate a battlecard with positioning, strengths, weaknesses, and…
Create beautiful HTML slide decks for product presentations. Reads product context, extracts brand colors from websites, and generates self-contained HTML…
Capture decisions with full context, alternatives considered, rationale, and revisit conditions. Auto-links to related PRDs, strategy docs, and OKRs.
Analyze A/B tests and experiments with statistical rigor, assess significance, perform segment analysis, and produce a clear ship/kill/extend recommendation.
Analyze customer feedback from any source. Categorize by theme, frequency, and severity. Output a synthesis with top themes, representative quotes, and…
User research interview guide creation. Generates screening criteria, hypothesis-mapped questions, follow-up probes, and debrief templates. Includes guidance…