The Agent Skills Directory

[COMMAND_EXECUTION]: The skill frequently executes shell commands using the Bash tool. It specifically invokes a local CLI named gsd-sdk and runs Node.js scripts located in $HOME/.claude/get-shit-done/bin/ (e.g., gsd-tools.cjs). This is the primary method for most of its 67 commands, such as gsd:graphify, gsd:config, and gsd:workstreams.
[REMOTE_CODE_EXECUTION]: The gsd:update command is designed to check for and install updates for the GSD toolset using npm (npm i -g get-shit-done). This involves downloading and executing code from the public npm registry at runtime.
[EXTERNAL_DOWNLOADS]: Several commands, including gsd:ai-integration-phase, gsd:plan-phase, gsd:sketch, and gsd:spike, are granted access to WebSearch and WebFetch tools to retrieve technical documentation and research data from the internet.
[DATA_EXFILTRATION]: While the skill accesses project files and potentially sensitive environment variables (noted in gsd:forensics), it does not appear to send this data to untrusted external domains. Its network activity is confined to research-related tools and the npm registry for updates.
[NO_CODE]: The skill package itself contains only instruction manifests (Markdown files with YAML frontmatter) and does not ship with the actual executable logic, which is expected to be installed separately in the user's home directory.

gsd-commands