pp-wikipedia
Pass
Audited by Gen Agent Trust Hub on May 8, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the agent to download and install CLI tools from repositories on GitHub and NPM associated with the author's workspace.
- [REMOTE_CODE_EXECUTION]: Installation instructions utilize npx and go install to fetch and install code from remote sources.
- [COMMAND_EXECUTION]: The skill allows the agent to execute a binary tool with an option to write output directly to the local file system via a --deliver file: parameter.
- [DATA_EXFILTRATION]: The provided CLI tool includes a --deliver webhook: feature that enables the agent to send data to arbitrary external webhooks.
- [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by combining external data ingestion with powerful capabilities:
- Ingestion points: Data is retrieved from the Wikipedia REST API, including article text and HTML (SKILL.md).
- Boundary markers: No specific delimiters or safety instructions are provided to help the agent distinguish article content from instructions.
- Capability inventory: The skill uses shell access, local file writing, and network POST requests (SKILL.md).
- Sanitization: No mechanisms are described for sanitizing or validating retrieved article content before the agent processes it.
Audit Metadata