Installs into .claude/skills of the current project.
Are you the author of Sniff Report?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/neversight-sniff-report)
---
name: sniff-report
description: Show the results from the last sniff scan.
user-invocable: false
---
# /sniff-report - Show last scan results
Display findings from the most recent scan.
This skill is invoked by the `sniff` skill's own flow, not on its own. The `user-invocable: false` frontmatter key is honored by Claude Code only, so in other agents this file may show up as a standalone command; treat it as a step inside a sniff run rather than an entry point.
## What to do
1. Use the unified `sniff` MCP tool with:
- `mode`: `"report"`
- `rootDir`: the current project's absolute path
- `format`: 'summary' for readable output, 'json' for raw data
2. If the unified `sniff` tool is not available, use legacy `sniff_report` with the same rootDir and format.
3. Present the summary to the user:
- Total findings by severity
- Top issues with file locations
- Any corroborated findings (confirmed in source + browser)
## Example
```
User: /sniff-report
You: Here are the results from your last scan:
**Summary:** 12 findings (1 critical, 3 high, 6 medium, 2 low)
**Top issues:**
1. `src/routes.ts:14` - CRITICAL: Hardcoded API key
2. `src/handler.ts:42` - HIGH: debugger statement
3. `README.md:28` - HIGH: Broken internal link
Reports saved to:
- .sniff/reports/report.html
- .sniff/reports/report.json
```
