Coding Lead
ReviewAudited by ClawScan on May 1, 2026.
Overview
This instruction-only coding skill openly enables code editing, command execution, Claude ACP delegation, and project context memory, which are powerful but aligned with its stated coding-execution purpose.
Install this only if you want an agent to perform implementation work on your projects. Use version control, review planned commands and diffs, keep ACP prompts and .openclaw context files free of secrets, and verify that any optional helper tools it calls are trusted.
Findings (4)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
A coding agent using this skill can modify project files and run local commands, so mistakes could change code, migrations, or local environment state.
The skill explicitly allows direct file reading, writing, editing, and command execution as part of its coding workflow.
- **Simple**:当前会话内直接 `read/write/edit/exec`
Use it in version-controlled project directories, review plans and diffs, and require explicit confirmation for destructive or environment-changing commands.
Behavior may depend on the provenance and safety of other installed OpenClaw skills or CLI tools.
The skill optionally relies on separately installed local tools and scripts that are not part of this artifact package.
**smart-agent-memory** | `node ~/.openclaw/skills/smart-agent-memory/scripts/memory-cli.js stats` ok? | `memory_search` + manual `.md` writes`
Only use this with trusted installations of acpx, qmd, and smart-agent-memory, and keep those tools updated from known sources.
Project context, prior decisions, or lessons may be reused across coding sessions, which can preserve useful continuity but may also carry stale or sensitive information forward.
The workflow creates persistent project context files and sets retention-style limits for active and archived context.
Context 文件治理 - active 上限:**60** - 生命周期总窗口:**100**(active + archive) - 推荐文件名:`<project>/.openclaw/context-<task-slug>.md`
Keep context files scoped to the task, avoid adding secrets, periodically clean old entries, and verify recalled context before relying on it.
Task prompts and project context may be handed to another coding runtime, and that runtime may read files in the configured project workspace.
The skill routes medium and complex coding work through Claude ACP or acpx when available, which is an explicit agent-to-agent/provider execution path.
1. `sessions_spawn(runtime:"acp")` 2. `acpx claude exec` 3. direct execution
Confirm the intended project cwd and context file contents before ACP runs, and do not include secrets or unrelated private files in prompts or context files.