Dev Project Tracker
v1.0.0按需求/版本维度管理项目全生命周期。创建需求目录、记录日常进展、追踪问题、管理变更、自动蒸馏归档。 触发词：创建需求目录、新建项目、项目进展、记录进展、项目归档、蒸馏项目、查看项目、项目状态、 "新需求"、"建个项目目录"、"记一下进展"、"这个需求上线了"。 适用于任何按版本/需求迭代的软件项目。
⭐ 0· 258·2 current·2 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 align with what is present: templates, README/WORK_LOG/ISSUES/CHANGES files and a small scanner script. Nothing requires cloud credentials or unrelated system access.
Instruction Scope
SKILL.md directs the agent to create/read/write files under projects/ and to update 'memory/YYYY-MM-DD.md' and the 'memory_store' (agent long-term memory). This is within the tracker purpose but worth noting: instructions both say 'do not mix personal logs' and also say '同步更新个人日志 memory/...', which is ambiguous. Also the scan script reads any README.md under the supplied projects root — ensure the configured projects root is correct so it doesn't scan unintended directories.
Install Mechanism
No install spec (instruction-only) and only one small shell script included. No downloads or extracted archives; low install risk.
Credentials
No environment variables, no credentials, and no external endpoints. The only sensitive action is writing key decisions to the agent's memory_store — expected for a tracker but has privacy implications the user should consider.
Persistence & Privilege
always is false and the skill does not request persistent system-wide privileges or modify other skills. Autonomous invocation is allowed by default (normal).
Assessment
This skill appears internally consistent, but before enabling it you should: (1) confirm where 'projects/' and 'memory/' will live (pass an explicit projects root when running the scanner) to avoid scanning or writing outside the intended workspace; (2) review your agent's memory retention/privacy settings because the skill instructs storing key decisions in memory_store; (3) test the included script in a safe directory — the script uses date -j -f (macOS-style) which may fail on some Linux systems; (4) back up important data before enabling automated archiving/cleanup; and (5) if you want stricter limits, run the skill in a constrained environment or restrict the agent's filesystem permissions so it can only access the intended project directories.Like a lobster shell, security has layers — review code before you run it.
latestvk97ctdatskxsay1jbfrrbcbxbx838b5s
License
MIT-0
Free to use, modify, and redistribute. No attribution required.