The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary function includes an automated synthesis pipeline that processes potentially untrusted external data. In synthesize-research-report.md and the accompanying phase instructions, the agent is directed to read and analyze raw interview transcripts, notes, and summaries provided by the user.
Ingestion points: External research data enters the context through files referenced in the synthesis pipeline.
Boundary markers: The instructions lack specific guidance on implementing boundary markers or filtering to prevent the agent from following instructions that might be embedded within the transcripts themselves.
Capability inventory: The agent has the ability to execute shell commands via the cookiy CLI, perform network operations to the vendor API, and write analysis files locally.
Sanitization: No explicit sanitization of the input text is described in the processing methodology.
[EXTERNAL_DOWNLOADS]: The skill's setup and operations rely on the installation of the @cookiyai/cookiy-cli Node.js package from a public registry (found in README.md and cookiy.md). This tool is provided by the skill's author, Cookiy AI.
[COMMAND_EXECUTION]: The skill uses shell command execution to interact with its platform. It frequently invokes the cookiy CLI to manage studies, handle recruitment, and check billing status.
[DATA_EXFILTRATION]: The skill transmits research-related data and metadata to the vendor's API at s-api.cookiy.ai. This includes study parameters, survey designs, and potentially base64-encoded image data uploaded via the study upload command.

user-research-cookiy