research

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process data from external, untrusted sources including job boards, social media, and industry forums, which presents a surface for malicious instruction injection. \n
  • Ingestion points: SKILL.md and reference/market.md instruct the agent to crawl LinkedIn, Glassdoor, Indeed, Reddit, and various industry forums. \n
  • Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the research templates. \n
  • Capability inventory: The agent has implicit network access for data retrieval and explicit instructions to write session results to the local filesystem (SKILL.md). \n
  • Sanitization: There are no instructions for sanitizing or validation of content retrieved from external sources before processing. \n- [COMMAND_EXECUTION]: The skill involves using external tools and commands to perform scraping and discovery of MCP servers, which could be leveraged to execute arbitrary logic. \n- [DATA_EXFILTRATION]: The skill provides instructions to read sensitive local configuration files such as claude_desktop_config.json for server discovery and writes execution metrics to the ~/.claude/skill-analytics/ directory, which involves accessing hidden directories used by the agent host environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 09:56 AM
Security Audit — agent-trust-hub — research