The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: Suspicious dependency usage detected.\n
The script scripts/ticktick_cli.py lists httpxyz>=0.42.1 as a dependency in its inline metadata block.\n
The file scripts/ticktick_api_client.py imports httpxyz and uses it to perform network operations, mimicking the API structure of the legitimate httpx library.\n
This name is a likely typosquat of the well-known httpx library, which is a common technique used to distribute malicious code through package registries.\n- [EXTERNAL_DOWNLOADS]: The skill attempts to install and load a non-standard third-party package.\n
Installation is triggered via the uv tool metadata in scripts/ticktick_cli.py.\n
The package source for httpxyz is unverifiable and mimics a trusted library name to evade detection.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the TickTick/Dida365 API which could contain malicious instructions.\n
Ingestion points: Data enters the agent's context through commands like project list, project data, task get, and task filter in scripts/ticktick_cli.py.\n
Capability inventory: The skill has the ability to execute shell commands (e.g., rm for cleanup), perform network requests to TickTick APIs, and manage sensitive authentication tokens stored locally.\n
Boundary markers: There are no instruction delimiters or specific warnings to the agent to ignore instructions embedded within task content or titles.\n
Sanitization: External content retrieved from the API is parsed directly into Pydantic models and displayed to the agent without escaping or filtering.

ticktick-cli