The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent to download and install CLI and MCP binaries from the vendor's GitHub repository (github.com/mvanhorn/printing-press-library) and via NPM (@mvanhorn/printing-press-library).
[COMMAND_EXECUTION]: Instructions include executing go install and npx -y to install and run software from external registries, which involves the execution of remote code during the setup process.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data from the Product Hunt API, user comments, and public Atom feeds.
Ingestion points: The tool fetches launch details, taglines, descriptions, and user-generated comments from the Product Hunt platform (SKILL.md).
Boundary markers: No specific prompt delimiters or instructions to ignore embedded commands within the retrieved web content are provided to the agent.
Capability inventory: The skill possesses the ability to execute arbitrary SQL (sql command), write output to arbitrary file paths (--deliver file:<path>), and perform network POST requests (--deliver webhook:<url>).
Sanitization: There is no evidence of explicit sanitization or filtering of the external Product Hunt data before it is presented to the agent's context.
[DATA_EXFILTRATION]: The CLI includes a --deliver webhook:<url> feature that allows the agent to route command results directly to an external endpoint. Additionally, the feedback command can be configured to POST local data to a remote server.
[DYNAMIC_EXECUTION]: The skill provides a sql command that allows the execution of arbitrary, user-supplied SQL queries against a local SQLite database. It also supports dynamic file writing to user-specified paths via the --deliver file:<path> sink.

pp-producthunt