pp-scientific-consensus
Pass
Audited by Gen Agent Trust Hub on Aug 28, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill requires the installation of the
scientific-consensus-pp-clibinary to function. - Evidence: The documentation instructs the agent to run
npx -y @mvanhorn/printing-press-library install scientific-consensus --cli-onlyorgo install github.com/mvanhorn/printing-press-library/library/other/scientific-consensus/cmd/scientific-consensus-pp-cli@latestto set up the necessary tools. - [DATA_EXFILTRATION]: The tool includes a delivery mechanism that allows command output to be sent to arbitrary external URLs.
- Evidence: The
--deliver webhook:<url>feature POSTs the output body to a user-defined URL, which represents a potential data egress point if not monitored. - Additional Evidence: The
feedbackcommand can be configured to automatically POST local feedback entries to a remote endpoint via theSCIENTIFIC_CONSENSUS_FEEDBACK_ENDPOINTenvironment variable. - [INDIRECT_PROMPT_INJECTION]: The skill processes content from various external scholarly databases, which may include untrusted text that could influence agent behavior.
- Ingestion points: Data is ingested from PubMed, OpenAlex, Crossref, and Europe PMC (SKILL.md).
- Capability inventory: The skill has the ability to execute shell commands (in agent mode), write to the local file system (Excel reports and local sinks), and perform network POST requests (webhooks).
- Boundary markers: None identified in the instructions to prevent the agent from obeying instructions embedded in the retrieved scientific abstracts or metadata.
- Sanitization: No explicit sanitization or filtering of the external API content is mentioned.
Audit Metadata