skills/first-fluke/mapple/oma-pdf/Gen Agent Trust Hub

oma-pdf

Warn

Audited by Gen Agent Trust Hub on Jun 18, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches the installation script for the 'uv' tool runner from Astral's official domain (astral.sh).
  • [COMMAND_EXECUTION]: Instructs the agent to interpolate user-provided passwords directly into shell commands (e.g., opendataloader-pdf -p "{password}"), which creates a risk of command injection if the input contains shell metacharacters.
  • [REMOTE_CODE_EXECUTION]: Recommends the use of the curl | sh pattern to install the 'uv' utility during the error recovery phase.
  • [COMMAND_EXECUTION]: Executes several external CLI tools, including opendataloader-pdf and mdformat, via the uvx runner to process local filesystem data.
  • [PROMPT_INJECTION]: Vulnerable to indirect prompt injection as the skill extracts and processes text content from untrusted PDF documents.
  • Ingestion points: PDF file content read from the local filesystem (resources/execution-protocol.md).
  • Boundary markers: None. The extracted text is processed and converted directly.
  • Capability inventory: Execution of shell commands via uvx and local file write operations (SKILL.md).
  • Sanitization: No evidence of sanitization, escaping, or filtering of extracted PDF content before it is used or written to output files.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jun 18, 2026, 12:59 PM
Security Audit — agent-trust-hub — oma-pdf