pp-amplitude

Pass

Audited by Gen Agent Trust Hub on Aug 20, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill facilitates the installation of the amplitude-pp-cli tool from external sources controlled by the author.
  • Installation occurs via npx -y @mvanhorn/printing-press-library and go install github.com/mvanhorn/printing-press-library/library/marketing/amplitude/cmd/amplitude-pp-cli@latest.
  • [REMOTE_CODE_EXECUTION]: The skill downloads and subsequently executes external binaries to perform analytics operations.
  • [COMMAND_EXECUTION]: The skill acts as a shell wrapper for the amplitude-pp-cli and amplitude-pp-mcp tools, allowing the agent to execute complex CLI commands.
  • [DATA_EXFILTRATION]: The CLI tool implements a --deliver webhook:<url> parameter, enabling query results to be POSTed to arbitrary external HTTP endpoints.
  • While documented as a data routing feature, this capability allows for automated data transfer outside the controlled environment.
  • [PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection due to its handling of external API data.
  • Ingestion points: Data is ingested from the Amplitude API via the users, usersearch, and event-exports commands in SKILL.md.
  • Boundary markers: No delimiters or instructions are provided to the agent to treat ingested data as untrusted or to ignore embedded instructions.
  • Capability inventory: The agent has access to shell command execution and external webhook delivery sinks.
  • Sanitization: There is no evidence of sanitization or rigorous validation for the data returned from the Amplitude API before it is processed.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 20, 2026, 08:30 AM
Security Audit — agent-trust-hub — pp-amplitude