SDD Brainstorming
v1.0.0在实现任何新功能、构建组件、添加功能或修改行为之前必须使用。用于在实现前探索用户意图、需求和设计。
⭐ 0· 137·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description (SDD brainstorming and design) align with the instructions: the SKILL.md only asks the agent to inspect a workspace configuration, read/write spec documents under {workspace}/spec/, ask clarifying questions, propose designs, and optionally launch image generation. There are no unrelated credentials, binaries, or installs requested.
Instruction Scope
Instructions explicitly read and write files under the configured workspace (.sdd-workspace and {workspace}/spec/...), scan existing feature folders, and read global constraints/docs; these actions are appropriate for a design helper. Note: the skill also launches a background subagent to call /gen-image for diagram generation, which will transmit design text/specs to whatever image-generation capability is invoked — confirm you are comfortable with that data leaving the local workspace.
Install Mechanism
No install spec and no code files — instruction-only skill (lowest install risk). Nothing is downloaded or written to disk by an installer.
Credentials
No environment variables, credentials, or external config paths are required. Access to workspace files is declared and proportional to the skill's purpose. The skill does read the agent/system prompt to check the model, which is consistent with its stated 'model check' step.
Persistence & Privilege
always is false and the skill does not request system-wide changes. It writes design docs to the specified workspace (expected). It can invoke other skills (normal); the only noteworthy behavior is launching a background subagent to generate images, which increases runtime activity but is coherent with the design purpose.
Assessment
This skill is internally consistent with its stated goal. Before installing or using it: (1) ensure your workspace (.sdd-workspace and files under {workspace}/spec/) does not contain sensitive secrets you wouldn't want read or sent to external services; the skill will read and write spec files and scan commits. (2) Confirm your platform's /gen-image (or equivalent image-generation skill) privacy/retention policy because the skill will launch a background subagent to generate diagrams and that will send specification content to the image generator. (3) The skill explicitly forbids performing implementation actions before user approval (and it also says not to auto git commit) — verify your agent enforces that HARD-GATE. If you need additional assurance, inspect the workspace contents the skill will access and run a test on a non-sensitive repo first.Like a lobster shell, security has layers — review code before you run it.
latestvk97cpmbs1mvvvdzrptm2nj1a91832w66
License
MIT-0
Free to use, modify, and redistribute. No attribution required.