pp-scientific-consensus
Pass
Audited by Gen Agent Trust Hub on Aug 6, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill facilitates the installation of the 'scientific-consensus-pp-cli' tool from the author's own repositories, specifically using 'go install' for GitHub modules under 'github.com/mvanhorn' and 'npx' for the '@mvanhorn/printing-press-library' package.
- [DATA_EXFILTRATION]: The CLI tool includes a '--deliver webhook:' parameter that allows command output to be automatically POSTed to an external URL. Furthermore, it supports the 'SCIENTIFIC_CONSENSUS_FEEDBACK_ENDPOINT' environment variable and '--send' flag to transmit locally collected feedback data to a remote server.
- [PROMPT_INJECTION]: The skill processes untrusted scholarly content from external sources, presenting an indirect prompt injection surface. * Ingestion points: Research paper metadata and summaries from PubMed, OpenAlex, Crossref, and Europe PMC. * Boundary markers: The instructions lack explicit delimiters to separate external data from system prompts. * Capability inventory: Includes shell command execution and the ability to write to network or file sinks via the delivery feature. * Sanitization: No content validation or escaping of the retrieved research data is described.
- [COMMAND_EXECUTION]: The skill functions as a wrapper for shell-based CLI execution. It includes a 'which' command that processes natural language input to determine the appropriate subcommand, which involves passing user-supplied strings to the underlying binary.
Audit Metadata