skills/mukul975/anthropic-cybersecurity-skills/extracting-browser-history-artifacts/Gen Agent Trust Hub
extracting-browser-history-artifacts
Pass
Audited by Gen Agent Trust Hub on Apr 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions include shell commands for mounting forensic disk images and performing batch extraction using the sqlite3 command-line utility.
- [DATA_EXFILTRATION]: The skill targets sensitive user data stored in browser profile directories.
- Evidence: Accesses Chrome and Edge History, Cookies, and Login Data files in agent.py and SKILL.md.
- Evidence: Accesses Firefox places.sqlite and cookies.sqlite databases in agent.py and SKILL.md.
- [PROMPT_INJECTION]: The skill ingests and processes untrusted data from browser artifacts, such as page titles and URLs, creating an indirect prompt injection surface.
- Ingestion points: agent.py reads from Chrome's History and Firefox's places.sqlite databases.
- Boundary markers: None present; extracted data is processed directly without delimiters or safety warnings.
- Capability inventory: The skill can read system files and write to the local file system using sqlite3 and csv standard libraries.
- Sanitization: No sanitization or validation of the extracted strings (URLs/titles) is performed.
- [EXTERNAL_DOWNLOADS]: The skill suggests installing the pyhindsight utility, a known third-party Python package for browser forensics.
Audit Metadata