The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions direct the agent to execute curl commands in the shell to interact with the up!link API for file uploads and URL ingestion.
[DATA_EXFILTRATION]: The skill is designed to send local files (-F "file=@<path>") to an external server. The destination host is dynamically determined from the UPLINK_API_KEY or UPLINK_BASE_URL environment variables, allowing data to be transmitted out of the local environment.
[INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data by ingesting remote URLs and uploading local files, which could lead to instructions being introduced into the agent's context.
Ingestion points: The skill ingests local file paths via the /api/upload endpoint and remote URLs via the /api/ingest-url endpoint as described in SKILL.md.
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands within the ingested content are present.
Capability inventory: The skill possesses the capability to read local filesystem data and perform network operations using curl.
Sanitization: There is no evidence of sanitization, filtering, or validation of the content being uploaded or ingested before it is processed.

uploading-with-uplink