ara-research-manager

Pass

Audited by Gen Agent Trust Hub on Jun 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes shell commands (mkdir -p) to initialize the directory structure for its research artifacts. This is a routine operation for its stated function of managing project documentation.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it ingests and processes untrusted data from the conversation history.
  • Ingestion points: Extracts information from the full conversation history and existing project files in the ara/ directory.
  • Boundary markers: The instructions do not define specific delimiters or warnings to ignore instructions embedded in the processed history.
  • Capability inventory: Accesses the file system to read and write Markdown/YAML files and create directories via shell commands.
  • Sanitization: There is no evidence of sanitization or filtering of the ingested conversation data before it is rewritten into project files.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 15, 2026, 01:11 PM
Security Audit — agent-trust-hub — ara-research-manager