O
openclaw-agent-optimize
About openclaw-agent-optimize
A skill for tuning OpenClaw workspaces through cost-aware model routing, parallel delegation, and leaner context. It guides operators through audits and guarded optimization without making persistent changes absent approval.
Key features
- 01Audit rules and memory
- 02Review tiered model routing
- 03Apply progressive context disclosure
- 04Parallelize independent work
- 05Add heartbeat and budget safeguards
Use cases
- 01Reviewing an OpenClaw workspace for optimization
- 02Reducing context from stale sessions
- 03Planning safer heartbeat changes
Install
Copy the command and run it in a terminal.
$
npx skills add https:/ / github.com/ phenomenoner/ openclaw-agent-optimize --skill openclaw-agent-optimizeAfter installing, type/openclaw-agent-optimizein Claude Code to confirm it loaded.
Bundled files
References and scripts installed along with SKILL.md.
README.mdreferences/cron-optimization.mdmemory-patterns.mdcontinuous-learning.mdoptimization-playbook.mdcontext-management.mdmodel-selection.mdheartbeat-watchdog-pattern.mdagent-orchestration.mdheartbeat-optimization.mdsafeguards.mdDetails
From the project READMEOpenClaw Agent Optimization
Use this skill to tune an OpenClaw workspace for cost-aware routing, parallel-first delegation, and lean context.
Safety Contract (must follow)
- Treat this skill as advisory by default, not autonomous control-plane mutation.
- Never mutate persistent settings (e.g.,
config.apply,config.patch,update.run) without explicit user approval. - Never create/update/remove cron jobs without explicit user approval.
- If an optimization reduces monitoring coverage, present options (A/B/C) and require the user to choose.
- Before any approved persistent change, show: (1) exact change, (2) expected impact, (3) rollback plan.
Workflow (concise)
- Audit rules + memory: ensure rules are modular/short; memory keeps only restart-critical facts.
- Model routing: confirm tiered routing (light / mid / deep) matches live config.
- Context discipline: apply progressive disclosure; move large static data to references/scripts.
- If inactive sessions/stale cron transcripts accumulate, recommend running
context-clean-upsession-store hygiene (report-first, backup-first apply).
- If inactive sessions/stale cron transcripts accumulate, recommend running
- Delegation protocol: parallelize independent tasks; use isolated sub-agents for long/noisy work.
- Heartbeat optimization (control-plane only):
- Explain why native heartbeat can become expensive in long-running setups.
- Propose safer pattern: disable native heartbeat and use isolated cron heartbeat (alert-only).
- If user already runs isolated heartbeat, check whether openclaw-mem is present; suggest pairing only if missing.
- Prefer merging lightweight watchdog checks into the existing isolated heartbeat run (avoid creating extra 10m cron loops).
- Offer profiles A/B/C if changing coverage.
- Safeguards: add anti-loop + budget guardrails; prefer fallbacks over blind retries.
- Execution gate: if user approves changes, apply the smallest viable change first, then verify and report.
References
references/optimization-playbook.mdreferences/model-selection.mdreferences/context-management.mdreferences/agent-orchestration.mdreferences/cron-optimization.mdreferences/heartbeat-optimization.mdreferences/heartbeat-watchdog-pattern.mdreferences/memory-patterns.mdreferences/continuous-learning.mdreferences/safeguards.md
Comments
Markdown supported