PDF Processing
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: HIGHPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- Indirect Prompt Injection (HIGH): The skill extracts text from untrusted external files, creating a significant attack surface for indirect prompt injection.
- Ingestion points: The
parser.parse("./my_file.pdf")call inSKILL.mdingests external data into the skill's workflow. - Boundary markers: Absent. There are no delimiters or instructions provided to the agent to treat the extracted text as untrusted data rather than instructions.
- Capability inventory: The extracted
full_textis aggregated into the agent's context, where it can influence downstream reasoning and decision-making. - Sanitization: Absent. No filtering or validation is performed on the parsed markdown content.
- Unverifiable Dependencies (MEDIUM): The skill requires the installation of the
llama_cloud_servicesPython package. - Source: PyPI.
- Status: The package is maintained by the LlamaIndex (run-llama) organization, which is not currently on the provided list of Trusted External Sources.
Recommendations
- AI detected serious security threats
Audit Metadata