The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external data from Pinterest search results, which is a common vector for indirect prompt injection attacks.
Ingestion points: Data enters the agent's context through the output of the postplus research run command, typically saved to result.json as specified in SKILL.md.
Boundary markers: The instructions do not define specific delimiters or warnings for the agent to ignore instructions that might be embedded within the search result titles or metadata.
Capability inventory: The skill allows the execution of CLI commands (postplus research run, postplus doctor) and local file writes (--output <result.json>).
Sanitization: The skill describes a normalization process to a fixed schema ({ image_url, pin_url, title }), which provides some structure but does not explicitly filter for malicious instructions in the text fields.
[COMMAND_EXECUTION]: The skill relies on the execution of the postplus CLI tool to perform its primary function.
Evidence: SKILL.md contains instructions for the agent to run postplus research run pinterest-search with user-supplied keywords. While this is the intended functionality using the author's own tools, it represents a command execution surface that requires the agent to interpolate user input into a shell command.

pinterest-search