The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes shell commands for environment discovery (such as jq and grep) and for executing test suites (like npm test, pytest, or cargo test). These actions are standard for development workflows and consistent with the skill's objective.
[PROMPT_INJECTION]: The skill ingests data from local configuration files to determine test commands, creating a surface for potential indirect instruction injection if those files are malicious. Ingestion points: Reads various project configuration files including package.json, Cargo.toml, Gemfile, go.mod, pyproject.toml, and Makefile within the user's workspace. Boundary markers: The skill uses predefined markdown templates in TASK_TEMPLATES.md to structure its output, though it does not explicitly use delimiters for the data read from configuration files. Capability inventory: Capabilities include shell command execution (git, package managers, and test runners) and local file system writes (generating task list files in the /tasks/ directory). Sanitization: No specific sanitization or validation of the content extracted from configuration files is described prior to execution.

generate-tasks