setup
Pass
Audited by Gen Agent Trust Hub on Mar 13, 2026
Risk Level: SAFEDATA_EXFILTRATIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill facilitates the setup of an optional telemetry system that sends session metadata (event types and timestamps) to an external webhook. It defaults to the author's infrastructure (hq.yonatangross.com) but explicitly requests user consent and discloses the data types being transmitted.
- [COMMAND_EXECUTION]: Bash is used to perform project diagnostics and scan for manifests (e.g., 'wc -l', 'claude --version'). It also executes internal TypeScript configuration scripts using 'npx tsx' to generate hooks and settings files based on user preferences.
- [EXTERNAL_DOWNLOADS]: The integrations phase recommends the installation of the 'agentation-mcp' package from npm to enhance frontend visual debugging capabilities.
- [PROMPT_INJECTION]: The skill has an indirect prompt injection surface through its repository scanning functionality.
- Ingestion points: Manifest files (package.json, pyproject.toml) and source code patterns are read during Phase 1 for stack detection.
- Boundary markers: No specific delimiters or 'ignore' instructions are used when reading file contents.
- Capability inventory: The skill utilizes 'Bash', 'Write', and configuration update tools.
- Sanitization: Pattern detection is performed directly on the contents of scanned files without prior filtering.
Audit Metadata