linkedin-export
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The script
scripts/li_ingest.pyusessubprocess.runto interact with therlamaandollamaCLI tools. These operations are used to create, list, and manage local semantic search collections based on the parsed LinkedIn data. The commands use hardcoded arguments or internally defined variables, minimizing the risk of arbitrary command injection. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external LinkedIn GDPR export files (such as messages and profile summaries) which could potentially contain adversarial instructions designed to influence the behavior of the AI agent during search or analysis tasks.
- Ingestion points:
scripts/li_parse.pyreads data from LinkedIn ZIP archives and CSV files. - Boundary markers: Output files generated by
scripts/li_export.pyandscripts/li_ingest.pyuse Markdown headers (e.g.,CATEGORY: LinkedIn Messages) and separators (---) to structure the data, though they lack explicit instructions to the agent to disregard embedded commands. - Capability inventory: The skill possesses the capability to execute shell commands via
subprocess.run(scripts/li_ingest.py) and write files to the local filesystem (scripts/li_export.py). - Sanitization:
scripts/li_export.pyimplements asafe_filenamefunction to sanitize output filenames using regular expressions. Content from the LinkedIn data itself is not specifically sanitized before being displayed or ingested into search collections.
Audit Metadata