linkedin-contrarian-takes
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill suggests the installation of additional tools using the
npx skills addcommand, referencing packages hosted under the author's own repository (liftli-ai/linkedin-agent-skills). - [COMMAND_EXECUTION]: Instructions are provided for the user to execute terminal commands to integrate Model Context Protocol (MCP) servers and log in to the vendor's service. These commands (
claude mcp add,claude mcp login) interact exclusively with the vendor's domain (mcp.liftli.ai). - [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes external user topics and personal experience data.
- Ingestion points: User-provided chat input and localized configuration files (
.liftli/voice.mdand~/.liftli/voice.md). - Boundary markers: The instructions do not define explicit delimiters or 'ignore' warnings for the ingested data.
- Capability inventory: The skill possesses the ability to read local files and invoke tools like
strategist_briefingif available. - Sanitization: There is no mention of specific sanitization or filtering logic for the processed input data.
Audit Metadata