The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructions provide a command to download and execute a shell script from a remote vendor server: curl -fsSL https://paperclip.gxl.ai/install.sh | bash. This is used to install the paperclip CLI on the agent's environment. While the domain is associated with the skill's author (K-Dense Inc. / GXL), the practice of piping remote scripts directly to a shell carries inherent risks of arbitrary code execution.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process large volumes of third-party content, including biomedical literature, clinical trial protocols, and regulatory documents. This content constitutes untrusted external data that could contain adversarial instructions aimed at influencing the agent's behavior.
Ingestion points: External papers and filings are retrieved via paperclip search, grep, cat, and map commands.
Boundary markers: The skill relies on the agent's internal reasoning and specific instructions rather than strict data delimiters.
Capability inventory: The skill has access to the Bash tool and can write files to the local system.
Sanitization: There is no explicit sanitization of the retrieved text before it is presented to the agent.
[COMMAND_EXECUTION]: The skill makes extensive use of the Bash tool to interact with the paperclip binary and standard shell utilities like grep, sed, head, and tail. It also provides instructions for sourcing environment variables from .env files using the set -a; . ./.env; set +a pattern.
[DATA_EXFILTRATION]: The skill contains commands that allow for data egress, such as paperclip upload (uploading local files to the vendor platform), paperclip share (granting access to document folders), and paperclip fetch (downloading content from external URLs using browser cookies). The instructions provide specific guardrails, advising the agent to use these features only upon explicit user request and never for sensitive system directories.
[CREDENTIALS_UNSAFE]: The skill documentation describes the use of PAPERCLIP_API_KEY for authentication. It instructs the agent to store this key in a .env file and source it into the environment, which is a standard practice for local CLI tools but requires careful management to avoid accidental exposure or commitment to version control.

paperclip