The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The setup protocol in protocols/setup.md instructions the agent to clone a repository from github.com/TechTechWizard/claude-work-tools.git and execute a script named install.sh. This involves running unverified code from an external source on the user's system.\n- [PERSISTENCE]: The skill instructs the agent to help the user modify shell configuration files such as ~/.zshrc or ~/.bashrc to update the PATH environment variable. Modifying these files is a standard persistence mechanism that allows code to run automatically in future sessions.\n- [INDIRECT_PROMPT_INJECTION]: The skill fetches and processes ClickUp task descriptions and user comments (via clickup task and clickup comments in protocols/read-task.md). This external data could contain malicious instructions designed to manipulate the agent's behavior.\n
Ingestion points: External task data and comments are ingested through the clickup CLI.\n
Boundary markers: The skill relies on natural language instructions for the agent to summarize tasks in its own words, which provides a weak boundary against injection.\n
Capability inventory: The skill has access to the Bash and Write tools and uses a custom CLI tool for task management.\n
Sanitization: There is no evidence of programmatic sanitization or validation of the content retrieved from the ClickUp API.\n- [EXTERNAL_DOWNLOADS]: The skill downloads the claude-work-tools repository from a remote GitHub source during its setup phase.\n- [COMMAND_EXECUTION]: The skill makes extensive use of the Bash tool to run the clickup CLI, git, and other shell-based installation commands.

clickup