case-study-writing

Pass

Audited by Gen Agent Trust Hub on Apr 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the infsh CLI tool for search queries and running applications, as defined in the skill's allowed-tools configuration.
  • [EXTERNAL_DOWNLOADS]: The documentation references a remote GitHub URL (https://raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md) for installing the required CLI dependencies.
  • [REMOTE_CODE_EXECUTION]: The skill includes a Python script executed via infsh/python-executor to generate charts using the Matplotlib library. This involves runtime code execution to process data and save an image file (results-chart.png).
  • [PROMPT_INJECTION]: The skill contains a surface for indirect prompt injection (Category 8) by fetching data from external web search services.
  • Ingestion points: Untrusted data enters the agent context via the tavily/search-assistant, exa/search, and exa/answer tools mentioned in SKILL.md.
  • Boundary markers: Absent; there are no specific instructions or delimiters provided to the agent to distinguish between the skill's logic and the content returned from external searches.
  • Capability inventory: The skill has capabilities for shell execution (infsh), file writing (saving images via Matplotlib), and network access (facilitated by the research tools).
  • Sanitization: Absent; the skill does not specify any validation or filtering of the information retrieved from external web sources before using it in the case study generation process.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 16, 2026, 12:27 PM