The Agent Skills Directory

[COMMAND_EXECUTION]: The skill serves as a wrapper for the sau (Social Auto Upload) CLI tool. It defines a protocol for the agent to execute shell commands for Douyin login, cookie verification, and content publishing.
Evidence: Found in SKILL.md and references/cli-contract.md which detail the sau douyin command-line interface.- [EXTERNAL_DOWNLOADS]: The skill provides instructions for setting up the execution environment, which involves installing Python packages and browser binaries.
Evidence: references/runtime-requirements.md contains instructions for installing social-auto-upload and using patchright to download Chromium.- [PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes user-supplied content (titles, descriptions, tags) and interpolates them into command-line arguments.
Ingestion points: External data fields such as --title, --desc, and --note defined in references/cli-contract.md.
Boundary markers: Not explicitly enforced in the markdown instructions, placing the responsibility on the agent implementation.
Capability inventory: The agent can execute arbitrary CLI commands via the sau tool.
Sanitization: The scripts/examples/douyin_cli_template.py file demonstrates best practices by using shlex.quote to sanitize arguments before execution.

douyin-upload