The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: In references/output-formats/tick/about.md, the skill instructs the user to install the 'Tick' CLI tool by downloading a script from a remote URL and piping it directly into bash: curl -fsSL https://raw.githubusercontent.com/leeovery/tick/main/scripts/install.sh | bash. This is a dangerous pattern that executes unverified remote code.
[COMMAND_EXECUTION]: The skill constructs shell commands to create tasks via the tick CLI using descriptions parsed from external specification files. The instructions in references/output-formats/tick/authoring.md only require escaping double quotes. They fail to sanitize other shell-active characters like backticks (`) or subshell markers ($()), which could allow for command injection if a specification file contains malicious content.
[EXTERNAL_DOWNLOADS]: The skill references external binaries and installation paths for the 'Tick' tool in references/output-formats/tick/about.md, including Homebrew taps and Go installs from the author's GitHub repositories.
[PROMPT_INJECTION]: The skill handles untrusted data from specification files which are used to generate tasks and influence agent behavior. This creates a surface for indirect prompt injection (Category 8).
Ingestion points: Specification content provided to SKILL.md and read by agents in references/read-specification.md.
Boundary markers: No explicit delimiters or boundary warnings are provided in the prompt templates.
Capability inventory: Execution of shell commands via the tick CLI, local file manipulation (create/delete), and interaction with the Linear MCP service.
Sanitization: Extremely limited; the instructions only mention escaping double quotes for shell arguments.

technical-planning