software-developer
The software-developer skill enables Claude to function as an autonomous engineer by writing and modifying code, executing tests via shell commands, and managing git repositories. Use this skill when requesting code implementation, bug fixes, feature development, or general software engineering tasks that require exploration of existing codebases, validation through compilation or testing, and version control integration.
git clone --depth 1 https://github.com/clawdotnet/openclaw.net /tmp/software-developer && cp -r /tmp/software-developer/src/OpenClaw.Gateway/skills/software-developer ~/.claude/skills/software-developerSKILL.md
When asked to "write code", "fix a bug", "implement a feature", or act as a "developer": 1) Analyze the Request: - Identify the target files, languages, and expected outcomes. - If the codebase is unknown, use the `shell` or `read_file` tools to explore the workspace (`ls`, `find`, or read `README.md`). 2) Plan the Implementation: - Break down the task into smaller logical steps. - For complex changes, write a brief plan before executing. 3) Execution: - Use `write_file` or `shell` to modify code. - Always run the relevant compiler or test suite using the `shell` tool after making changes to verify they compile and pass. - Do not assume code works without validating it locally. 4) Version Control: - If requested, use the `git` tool to commit changes. - Write clear, descriptive commit messages. 5) Constraints: - Do not modify files outside the intended project scope. - Respect existing code style and architecture.
Organizes repository documentation and keeps new docs in the correct location.
Produce a daily news brief (AI/security/dev) with links, key takeaways, and action items.
Connects to databases, runs SQL queries, and analyzes datasets using code to provide actionable business insights.
Conducts comprehensive web research, synthesizes data from multiple sources, and produces detailed reports.
Triage inbox into “needs reply”, “FYI”, “urgent”, and “safe to archive”, with a short plan for each.
Safely operate Home Assistant by resolving targets first and validating state before writes.
Safely interact with MQTT topics with allow/deny policies and minimal payload risk.