anysite-cli
Pass
Audited by Gen Agent Trust Hub on Mar 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill operates by executing anysite CLI commands through the system shell (Bash). This includes API calls, dataset collection, and database management tasks.
- [EXTERNAL_DOWNLOADS]: The documentation instructs users to install the anysite-cli Python package from PyPI.
- [CREDENTIALS_UNSAFE]: The skill provides instructions for managing API keys and database credentials. While it demonstrates passing passwords directly in commands (e.g., anysite db add ... --password secret), it also correctly recommends using environment variables (e.g., --password-env PGPASS) to prevent secrets from appearing in command history.
- [PROMPT_INJECTION]: The skill provides a surface for indirect prompt injection via the processing of untrusted external web data.
- Ingestion points: Data is ingested from various external web platforms (LinkedIn, Instagram, Twitter, etc.) via API endpoints and from user-provided files (from_file) as seen in SKILL.md.
- Boundary markers: The provided documentation does not explicitly detail the use of boundary markers or instructions to ignore embedded commands when processing web data through LLM analysis commands.
- Capability inventory: The skill has the capability to write data to the local filesystem (Parquet, CSV, JSON), load data into external databases (PostgreSQL, ClickHouse, SQLite), and send data to remote webhooks, as described in references/dataset-guide.md.
- Sanitization: The skill documentation specifies that the transform.filter feature uses a safe expression parser rather than eval(), mitigating some risks of code injection during data processing.
Audit Metadata