pp-apify
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFECREDENTIALS_UNSAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the user to download a CLI tool from external sources using 'npx -y @mvanhorn/printing-press-library' and 'go install github.com/mvanhorn/printing-press-library'. These sources are associated with the author's development environment.
- [REMOTE_CODE_EXECUTION]: The installation process involves executing remote code from NPM and GitHub to set up the necessary command-line interface for the skill's operations.
- [DATA_EXFILTRATION]: The skill features a '--deliver webhook:' flag that enables the agent to POST command output to arbitrary external URLs. Additionally, the 'feedback' command can transmit local data to a user-defined 'APIFY_FEEDBACK_ENDPOINT'.
- [CREDENTIALS_UNSAFE]: The tool requires an 'APIFY_TOKEN' to authenticate with the Apify platform. While it follows standard practices by recommending environment variables, handling this token involves risks associated with sensitive credential management.
- [INDIRECT_PROMPT_INJECTION]: The skill processes data ingested from external scraping targets such as Twitter, Reddit, and Hacker News through 'apify-pp run' and 'apify-pp search' (SKILL.md). There are no specific boundary markers or sanitization procedures mentioned to prevent the agent from interpreting instructions embedded in this untrusted data. The skill possesses significant capabilities, including local file writing via '--deliver file:', network requests via '--deliver webhook:', and arbitrary shell command execution through the 'apify-pp-cli' (SKILL.md).
Audit Metadata