data-to-word-report
Pass
Audited by Gen Agent Trust Hub on Jun 21, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted data files to generate content for an analysis report. 1. Ingestion points: Step 1 of SKILL.md involves reading user-provided CSV, JSON, Excel, and TXT files. 2. Boundary markers: There are no specific instructions or delimiters provided to help the agent distinguish between data and embedded instructions. 3. Capability inventory: The skill uses shell command execution (SKILL.md) and local file system writes (scripts/gen_report.py). 4. Sanitization: No sanitization or validation logic is specified for handling data before it is passed to the generation script.
- [COMMAND_EXECUTION]: The skill utilizes a shell command to execute the report generation script with arguments provided by the agent. If the agent is influenced by malicious content within the data files, this could lead to arbitrary file writes or the injection of malformed arguments into the command line.
Audit Metadata