easy-ui-mcp: Skill for Claude Code

Role: HTML Report Renderer

You transform the output of the immediately preceding Stage 4 skill into a self-contained HTML report page using the project's report template.

Activation

Invoked by the Supervisor immediately after code-review, security-review, or blast-radius completes.

Skill({ skill: "html-report", args: "skill=code-review task=T001 branch=main" })

Workflow

1. Read the preceding skill output

The output of the just-completed Stage 4 skill is already in the current conversation context above. Read it carefully. Identify:

• All findings (each with severity: High / Med / Low / Info, file path, line range, description)
• The overall narrative / summary prose
• Any explicit risk signals, vulnerability classifications, or hardening recommendations

2. Score the three dimensions (0–100 integers)

Risk % (0 = no risk, 100 = critical risk)

• Count High findings: each adds ~15 pts
• Count Med findings: each adds ~7 pts
• Count Low findings: each adds ~2 pts
• Security vulnerabilities (from security-review or blast-radius) add extra weight
• Cap at 100

Code Quality % (0 = poor, 100 = excellent)

• Start at 90
• Subtract per High finding: −12
• Subtract per Med finding: −6
• Subtract per Low finding: −2
• Add back for positive signals (good test coverage, clear structure mentioned in narrative): +5 each
• Floor at 5

Adaptation Effort % (0 = no work, 100 = major rework required)

• Estimate total effort to address all findings
• 1–2 trivial fixes: ~10–20%
• 3–5 moderate fixes: ~30–50%
• Many fixes or architectural changes: ~60–90%

Overall health (badge):

• Derive from Risk score: 0–30 → green / "Healthy", 31–65 → yellow / "Needs Attention", 66–100 → red / "Critical"
• If Code Quality < 40, bump badge one level worse (green→yellow, yellow→red)

3. Build the findings rows

For each finding, produce one <tr> block. Severity class mapping:

• High → sev-high
• Med → sev-med
• Low → sev-low
• Info → sev-info