← Back to Skills Marketplace
Openclaw Optimizer
by
Jacob Ben-David
· GitHub ↗
· v1.19.0
· MIT-0
581
Downloads
0
Stars
4
Active Installs
5
Versions
Install in OpenClaw
/install openclaw-optimizer
Description
Use when: you want to optimize an OpenClaw setup (v2026.2.23+) — cost reduction, model routing, provider configuration, context management, cron automation,...
Usage Guidance
This skill appears coherent for optimizing and troubleshooting OpenClaw. Before using it: (1) inspect the included scripts (scripts/update-skill.sh and scripts/version-check.py) so you know what network or git actions they perform if you run them; (2) treat any proposed 'apply' CLI commands as manual operations — verify diffs and have backups (the SKILL.md already recommends backups); (3) be aware the skill will read local OpenClaw bootstrap/identity/config files during audits (this is expected), and some suggested fixes may include running system commands (e.g., launchctl, service restarts) or openclaw commands that can execute on nodes — only run those when you’re prepared; (4) if you plan to use update/commit helpers, ensure your git credentials and remote are what you expect; (5) if you want extra safety, run the audit mode (the skill's recommended 'Full audit (safe, no changes)') first and review all outputs before applying changes.
Capability Analysis
Type: OpenClaw Skill
Name: openclaw-optimizer
Version: 1.19.0
The skill is a highly powerful administrative framework for OpenClaw that manages sensitive configurations, credentials, and system persistence (cron, launchctl). It introduces a 'System Profile' feature (Section 11) that stores deployment-specific metadata, including IPs, SSH access details, and truncated secrets, with instructions for syncing this data via `scp`. Additionally, the 'Identity Optimizer' (references/identity-optimizer.md) provides a structured workflow for the agent to read and modify core personality (SOUL.md) and user preference (USER.md) files. While the skill includes a 'Safety Contract' and warns against malicious third-party skills, its broad capabilities—including self-modification via `update-skill.sh` and extensive network/filesystem access—create a significant attack surface for prompt injection and potential data exfiltration.
Capability Assessment
Purpose & Capability
Name/description (OpenClaw optimizer) match the contents: extensive CLI guidance, provider reference, identity audit, cron and troubleshooting docs. The skill includes helpers for version-awareness and explicit workflows to read and propose config changes — all expected for this purpose.
Instruction Scope
SKILL.md instructs the agent to read OpenClaw workspace files (SOUL.md, IDENTITY.md, AGENTS.md, USER.md), run OpenClaw CLI commands, and run local troubleshooting commands (e.g., launchctl getenv on macOS). Those actions are within scope for an optimizer but do involve reading local config and running system-level CLI commands. The skill's Safety Contract states changes will not be applied without explicit approval, which limits risk — nevertheless, any proposed 'apply' steps or script executions should be manually reviewed before running.
Install Mechanism
No install spec; this is instruction-first with two small helper scripts included. No external downloads or archive extraction are declared. Because scripts are present, review them before executing; but the package does not attempt to auto-install or pull binaries by itself.
Credentials
The skill does not require any environment variables or credentials in its metadata. The provider reference legitimately documents many provider env var names and shows example exports (e.g., ANTHROPIC_API_KEY, AWS_*), which is appropriate for a provider/configuration advisor. There is no unexplained request for unrelated secrets or system-wide credentials in the manifest.
Persistence & Privilege
always:false and no elevated platform privileges are requested. The skill includes user-initiated update tooling (update-skill.sh) that can commit/push if you run it explicitly — this is documented and not automatic. Autonomous invocation is allowed (normal default) but combined with the Safety Contract the skill claims to be advisory by default.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install openclaw-optimizer - After installation, invoke the skill by name or use
/openclaw-optimizer - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.19.0
openclaw-optimizer v1.19.0 — Major update for OpenClaw v2026.3.8+
- Expanded provider support from 29 to 40+ (see updated references/providers.md).
- Updated provider guidance with critical March 2026 warnings on Google Gemini/Anthropic bans and provider best practices.
- Added new “openclaw backup create” backup layer and verification workflow.
- CLI and model table updated for new model/slug conventions (OpenAI, Gemini, etc).
- Troubleshooting, onboarding, and audit references refreshed to match OpenClaw v2026.3.8.
- Updated safety and backup contract sections to reflect new workflows and provider rules.
v1.16.1
openclaw-optimizer v1.16.1
- Updated SKILL.md: expanded documentation, including new provider tables, error troubleshooting references, backup best practices, and CLI-first audit/advisory workflow.
- Documented explicit version check/update mechanisms and clarified backup strategies.
- Added detailed quick start prompts and safety contract for user-driven change approval.
- Aligned guidance and references with OpenClaw v2026.2.26 and latest provider status (including Gemini API/OAuth warning).
v1.16.0
- Updated version alignment to OpenClaw v2026.2.26 and incremented skill version to 1.16.0
- Documentation in SKILL.md refreshed and reorganized: clarified supported providers, update workflow, safety contract, and backup strategy
- Enhanced instructions for provider configuration, troubleshooting, and quick-start prompts
- Improved visibility of advisory defaults and explicit-change requirements
v1.15.1
**Changelog for openclaw-optimizer v1.15.1**
- Added metadata and reference files for knowledge mapping, update logs, version tracking, and troubleshooting.
- Introduced scripts for version checking and updating the skill (`update-skill.sh`, `version-check.py`).
- Included system and documentation templates to streamline new reference additions.
- No user-facing workflow, logic, or behavior changes; all additions support maintainability and future-proofing.
v1.15.0
openclaw-optimizer v1.15.0 — Release aligned with OpenClaw v2026.2.26, focusing on enhanced version awareness, safety, and up-to-date provider support.
- Adds runtime and GitHub Actions-based version drift checks for OpenClaw compatibility, including user prompts for updating.
- Updates provider reference: supports all 29 model providers with full CLI instructions and notes on OAuth/API key risks.
- Clarifies backup strategy with three-layer coverage and explicit guidance for when manual backups are needed.
- Expands quick start prompt examples for audits, troubleshooting, provider setup, routing optimizations, cron jobs, and identity/personality reviews.
- Tightens safety contract: all changes are advisory by default and require explicit user approval with pre-change impact/rollback instructions.
Metadata
Frequently Asked Questions
What is Openclaw Optimizer?
Use when: you want to optimize an OpenClaw setup (v2026.2.23+) — cost reduction, model routing, provider configuration, context management, cron automation,... It is an AI Agent Skill for Claude Code / OpenClaw, with 581 downloads so far.
How do I install Openclaw Optimizer?
Run "/install openclaw-optimizer" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Openclaw Optimizer free?
Yes, Openclaw Optimizer is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Openclaw Optimizer support?
Openclaw Optimizer is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Openclaw Optimizer?
It is built and maintained by Jacob Ben-David (@jacob-bd); the current version is v1.19.0.
More Skills