Feature Task Planning
v1.0.0将技术方案拆解为可执行的开发任务清单，每个任务适配 TDD 流程。当用户说'拆任务'、'开始规划任务'、'技术方案定了，接下来怎么开发'、'帮我拆一下开发计划'等意图时触发。
⭐ 0· 71·1 current·1 all-time
byp1866@cping6
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description (拆任务 / task planning) match the SKILL.md: it requires reading project spec files and producing a task plan. There are no unrelated environment variables, binaries, or services requested.
Instruction Scope
Runtime instructions are limited to reading repository docs (specs/PROJECT-CONTEXT.md, specs/features/{功能名称}*.md), using the provided assets template, producing a tasks document, and validating AC coverage. These file reads/writes are proportional and relevant to the stated goal; there are no instructions to call external endpoints, read arbitrary system files, or access secrets.
Install Mechanism
No install spec and no code files — instruction-only. No downloads, package installs, or archive extraction are present.
Credentials
The skill requests no environment variables, credentials, or config paths. All declared preconditions and file accesses align with the planner's purpose.
Persistence & Privilege
always is false and model invocation is normal default. The skill intends to write generated docs to specs/features/{功能名称}_任务规划.md — this is expected and scoped to the repo. It does not request persistent elevated privileges or modify other skills' configs.
Assessment
This skill is instruction-only and appears to only read project spec files and write a task-plan markdown under specs/features/..., which is expected for a task-planning helper. Before installing or enabling it: 1) Confirm you are comfortable with the agent having repository read/write access to the specs/ paths (it will read specs/PROJECT-CONTEXT.md and specs/features/... and write specs/features/{功能名称}_任务规划.md). 2) Ensure sensitive data is not stored under those spec paths (the skill does not need secrets). 3) Review the generated file before committing to source control. 4) If you allow autonomous agent invocation generally, remember this skill can be invoked by the agent when the user intent matches—this is normal but means any file writes should be reviewed. If you want, restrict the agent's repo access or run the skill in a sandboxed environment the first time.Like a lobster shell, security has layers — review code before you run it.
latestvk976bw9btb7jn9bqhhgvtx2ym183kxc2
License
MIT-0
Free to use, modify, and redistribute. No attribution required.