Content
57%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 delivers genuinely actionable security content — detection patterns, framework-specific checks, secure alternatives, tooling commands, and a prevention checklist — but is undermined by boilerplate filler sections, a long re-implementation of knowledge Claude already has, and an all-in-one monolithic layout with no bundle files. The strongest fix is trimming the filler, moving the scanner implementation to scripts/, and adding an explicit verify-findings step.
Suggestions
Remove the circular 'Use this skill when', 'Do not use', 'Context', and 'Requirements' filler sections; the frontmatter description already carries trigger and scope information.
Replace the ~150-line scanner class with a compact vulnerable-pattern table (or move the implementation to a scripts/xss-scanner.ts bundle file), since Claude already knows these XSS patterns — keep the prevention checklist and secure-pattern examples inline.
Add an explicit workflow step for validating findings (e.g., confirm user-control reachability and filter false positives) between detection and report generation, since the naive indicator matching ('props', 'state', 'query') will over-flag.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body includes circular filler ("Working on xss vulnerability scanner for frontend code tasks or workflows", "Take the requirements, target, and scope from the user's request") and ~150 lines of TypeScript re-implementing XSS pattern detection (innerHTML, v-html, dangerouslySetInnerHTML) that Claude already knows. The core checklist and secure-pattern content is dense and useful, so it is above anchor 2's noticeably-verbose profile but clearly below anchor 4's efficient one. | 3 / 5 |
Actionability | Mostly executable guidance: concrete TypeScript scanner classes, runnable commands ("semgrep --config=p/xss --json", eslint-plugin-security install), and copy-paste secure-pattern snippets. Minor gaps hold it below anchor 5: detectEventHandlerIssues is called but never defined, findJavaScriptFiles is missing, fs is unimported, and "node xss-scanner.js --path=src" references a file that does not exist in the bundle. | 4 / 5 |
Workflow Clarity | Numbered sections 1-6 (detection, framework-specific, secure examples, tooling, reporting, prevention) plus an Output Format list give a topical sequence, but there is no ordered run-through of a scanning engagement and no validation checkpoints such as verifying findings or filtering false positives. Scanning is read-only, so no destructive-operation cap applies; this matches anchor 3 (sequence present, checkpoints missing or implicit). | 3 / 5 |
Progressive Disclosure | The skill is a single ~300-line file with no references/, scripts/, or assets/ bundle; ~200 lines of scanner and report-generator code are inlined in SKILL.md where they belong in a scripts/ file, with the checklist kept inline. Section headers provide real structure, so it fits anchor 3 (structure present but content that should be separate is inline) rather than anchor 2's unstructured wall; it cannot reach anchor 4-5 since nothing is split out. | 3 / 5 |
Total | 13 / 20 Passed |