Menu
Audit the Claude Code harness configuration for reliability, cost, and throughput via the harness-optimizer agent.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Scan installed skills to extract cross-cutting principles and distill them into rules — append, revise, or create new rule files.
Audit all installed OpenClaw skills for quality: content overlap, freshness, and uniqueness. Supports Quick Scan (changed only) and Full Stocktake modes.
Cluster MEMORY.md entries by theme and optionally generate SKILL.md drafts.
Export MEMORY.md to a timestamped markdown file for sharing or backup.
Import memory entries from a markdown export file, merging with MEMORY.md without duplicating.
Show current workspace memory: MEMORY.md contents, pending recalls, and index status.
| name | cmd_harness_audit |
| description | Audit the Claude Code harness configuration for reliability, cost, and throughput via the harness-optimizer agent. |
| user-invocable | true |
| origin | openclaw-mas |
| argument-hint | <project-path> |
Delegate to the harness-optimizer agent.
Include in the task payload:
First, reply to the user briefly to confirm you are delegating to harness-optimizer.
Then call sessions_spawn:
{
"agentId": "harness-optimizer",
"sessionKey": "harness-optimizer",
"task": "<user's full request and all relevant context — the agent cannot see this conversation>",
"runTimeoutSeconds": 0
}
After sessions_spawn returns, relay the result to the user. Do not output anything after the spawn call until the result arrives.