The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing untrusted data from social media platforms.
Ingestion points: Public note content, creator bios, and top-level comments are fetched via scripts/lingzao_client.py and processed by the agent.
Boundary markers: Absent. The instructions do not use delimiters or explicit 'ignore embedded instructions' warnings when presenting external content to the model.
Capability inventory: The skill can execute local shell commands (suggested curl operations in scripts/lingzao_client.py), perform network operations to its API backend, and write files to the local system via the generate-image command.
Sanitization: Content is summarized and analyzed directly without filtering for embedded instructions, though the CLI client uses shell quoting for command arguments.
[COMMAND_EXECUTION]: The skill uses local scripts and generates a CLI wrapper to manage its operations.
Evidence: scripts/setup.sh creates a wrapper at ~/.lingzao/bin/lingzao and scripts/configure.py is used to manage local API configuration with restricted file permissions (0600).
[EXTERNAL_DOWNLOADS]: The skill performs network operations to fetch updates and content data from vendor-controlled domains.
Evidence: scripts/lingzao_client.py fetches version information from assets-tian.midao.site and interacts with the API at lingzao.atian.vip. These are identified as legitimate vendor resources for the 'atian-create' author.

lingzao