pp-printing-press-rich
Pass
Audited by Gen Agent Trust Hub on Sep 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill's installation instructions download and execute code from a remote source.
- Evidence: README.md and SKILL.md instruct the user to run npx -y @mvanhorn/printing-press-library install printing-press-rich to install the required CLI and agent components.
- Context: The package belongs to the skill's author (mvanhorn) and is a necessary step for the skill's functionality.
- [COMMAND_EXECUTION]: The skill interacts with the local system by executing the printing-press-rich-pp-cli binary.
- Evidence: SKILL.md and internal/mcp/tools.go describe and implement the execution of CLI commands in response to user requests or tool calls.
- Context: This is the primary function of the skill, providing a bridge between the AI agent and the CLI tool.
- [DATA_EXFILTRATION]: The CLI supports routing command output to external URLs via a webhook sink.
- Evidence: helpers.go implements a --deliver webhook: feature that POSTs command results to a user-provided URL.
- Context: This is an intended feature for automating data workflows and does not appear to be used for unauthorized data harvesting.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from API responses and user inputs which are used to generate future agent instructions.
- Ingestion points: Data enters the agent context through user-provided queries (SKILL.md), external API responses (helpers.go), and local SQLite storage (tools.go).
- Boundary markers: Explicit delimiters or 'ignore embedded instructions' warnings are not used for these data sources.
- Capability inventory: The skill can execute shell commands (SKILL.md), write to configuration files (config.go), and perform network operations (tools.go).
- Sanitization: The skill employs cliutil.ScrubTerminal to clean output (helpers.go) and a custom SQL validator (validateReadOnlyQuery in tools.go) to prevent malicious queries from affecting the local database.
- [SAFE]: The SQL analysis tool includes dedicated security logic to ensure only read-only queries are executed, effectively mitigating risks associated with arbitrary SQL execution on the local store.
Audit Metadata