The Agent Skills Directory

[PERSISTENCE]: The skill provides instructions for the agent to modify the user's shell configuration file (~/.bashrc) to update the PATH environment variable, ensuring the downloaded binaries are executed in future sessions.
[REMOTE_CODE_EXECUTION]: The skill directs the agent to download Node.js binary archives from nodejs.org or npmmirror.com, extract them to a local directory, and execute the resulting binaries. This involves downloading and running third-party code on the host machine.
[EXTERNAL_DOWNLOADS]: Fetches binaries, checksum files, and configuration scripts from external domains including nodejs.org, npmmirror.com, and iqinghu.com for installation and setup.
[COMMAND_EXECUTION]: The skill makes extensive use of shell commands such as curl, tar, npm, and mkdir to perform system-level operations and environment configuration.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the user and interpolates it into the qhkit command line tool.
Ingestion points: Data enters the system via the uploadedImages (image paths) and prompt (text) fields in the JSON payload described in SKILL.md.
Boundary markers: None; the inputs are directly interpolated into the shell command strings.
Capability inventory: The skill performs network operations (API calls via qhkit), file system modifications, and shell command execution.
Sanitization: There is no evidence of input validation, escaping, or filtering to prevent malicious payloads from influencing the tool's behavior or the underlying system.

qinghu-image-variations