← Back to Skills Marketplace
247
Downloads
0
Stars
1
Active Installs
21
Versions
Install in OpenClaw
/install spectyra
Description
Optimize OpenClaw workflows and reduce LLM API costs. Runs locally to reduce repeated context, unnecessary steps, and token waste with no workflow changes re...
Usage Guidance
This skill is coherent: it installs a local helper and configures OpenClaw to route model calls to it. Key points to consider before installing: 1) The companion is a local HTTP service that will receive your full agent prompts (which may include secrets). Only run it if you trust the publisher; inspect the companion's source code or npm package (and its GitHub repo) before installing. 2) npm -g installs can run postinstall scripts; prefer auditing the package or installing in an isolated environment/container if unsure. 3) The skill will merge config into your agent to set spectyra as a model provider/default — if you don't want automatic routing, decline the config merge or remove the provider after testing. 4) The companion may later ask for upstream API keys (OpenAI/Anthropic/etc.) to forward requests — never provide sensitive credentials unless you trust the software. If you want higher assurance, review the @spectyra/local-companion code and repository (https://github.com/ImMrLucky/spectyra.git) before installing.
Capability Analysis
Type: OpenClaw Skill
Name: spectyra
Version: 1.0.25
The skill bundle is a configuration wrapper for 'Spectyra', a tool designed to optimize LLM costs by routing traffic through a local proxy. The files (SKILL.md, skill.json, config-fragment.json) define model providers pointing to a local endpoint (127.0.0.1:4111) and provide instructions to install the companion service via npm (@spectyra/local-companion). No evidence of malicious intent, data exfiltration, or obfuscation was found; the behavior is transparent and aligns with the stated purpose of local workflow optimization.
Capability Tags
Capability Assessment
Purpose & Capability
The skill name/description say it runs a local companion to optimize workflows and reduce token usage. The declared binary (spectyra-companion), the npm install package (@spectyra/local-companion), and the config-fragment that points models to http://127.0.0.1:4111 are all directly related to that purpose and proportionate.
Instruction Scope
SKILL.md instructs the user to globally install the companion and run spectyra-companion; it also installs a config merge that routes spectyra/* model IDs to the local server. These instructions are within scope, but they cause the agent to send full prompts (and any sensitive content) to a local HTTP service. The skill does not request unrelated files or environment variables, but the companion itself (not included here) will see agent traffic and may ask for upstream API keys or other configuration.
Install Mechanism
Installation is via a published npm package (@spectyra/local-companion) which is a common distribution mechanism. npm global installs can run lifecycle scripts and write binaries (spectyra-companion). No arbitrary download URLs or archive extraction are used in the skill metadata, but the companion package source is not included in the skill bundle so its contents and postinstall behavior are not visible here.
Credentials
The skill declares no required environment variables or secrets. The config-fragment uses a placeholder apiKey 'SPECTYRA_LOCAL' for local routing (not a secret). Note: the companion may later request upstream provider API keys via its UI or config, but that is outside the skill's declared requirements.
Persistence & Privilege
The skill's setup performs a config-merge that adds a local provider and sets spectyra/smart as the default model — this modifies agent configuration (expected for a local provider integration). 'always' is false. Because the change redirects model traffic to a local server, it increases the blast radius (the companion will receive prompts), so users should be aware before accepting the config merge.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install spectyra - After installation, invoke the skill by name or use
/spectyra - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.25
- No changes detected in this version.
- Version number updated, but all content and configuration remain the same.
v1.0.24
- Updated to OpenClaw metadata version 1.0.24.
- Improved description to clarify local savings and cost reduction without reducing output quality.
- Added installation instructions for installing Spectyra as an OpenClaw skill.
- Revised sections for clarity and streamlined setup guidance.
v1.0.23
- Updated description for improved clarity on optimization and cost savings.
- Added "version: 1.0.23" to OpenClaw metadata.
- Enhanced documentation with emphasis on reducing token waste and unnecessary steps.
- Clarified local running process and benefits with no workflow changes required.
v1.0.22
spectyra 1.0.22
- Updated description for clarity on installation and usage steps.
- Removed version pinning from OpenClaw integration metadata.
- No changes to features or functionality.
v1.0.21
- Shortened and simplified documentation for easier onboarding.
- Updated installation and run instructions for @spectyra/local-companion.
- Streamlined dashboard and model usage guidance.
- Updated OpenClaw compatibility version in metadata.
v1.0.20
- Updated OpenClaw compatibility to version 1.0.20.
- Added explicit disclaimer that SKILL.md cannot enforce runtime behavior; directs risk reviewers to companion source and README for verification.
- Clarified privacy and trust posture, emphasizing that all claims here are product descriptions and true guarantees require source inspection.
- Introduced an "Audit pointers" table with direct links to public source files covering companion runtime, telemetry, and auth pathways.
- Minor text and section adjustments for clearer risk review and separation of documentation from actual runtime behavior.
v1.0.19
- Updated to OpenClaw and companion version 1.0.19.
- Improved trust boundary explanation, clarifying explicit loopback routing and what stays local.
- Added recommended "npx" run command, reducing need for global installs.
- Made privacy and metric collection notes clearer and more concise.
- Streamlined description and dashboard access instructions.
v1.0.18
- Clarified that the official @spectyra/local-companion npm package is installed, emphasizing its origin and trust boundary.
- Added a new reviewers section explaining local traffic flow, data privacy, and security considerations.
- Updated OpenClaw version requirement to 1.0.18.
- Improved wording in the description and installation instructions for better transparency.
- No functional changes to companion usage or local optimization flow.
v1.0.17
- Added OpenClaw provider config version pin (1.0.17) under metadata.
- Clarified that provider config now uses `http://localhost:4111/v1` (instead of the numeric loopback) to avoid security scan “raw IP” warnings, while still reaching the same local companion endpoint.
v1.0.16
- Added a new "Privacy" section to clarify local handling of prompts, credentials, and anonymous product counters.
- Updated documentation to explain how to disable anonymous usage reporting via environment variable.
- No code changes; this update is documentation only.
v1.0.12
updated to show correct dashboard URL when launching local companion
v1.0.11
Spectyra 1.0.13 — Added savings details and optimization workflow overview.
- Updated dashboard section to highlight real-time local savings display.
- Added note on observed savings (60%–70%) from using Spectyra for LLM API calls.
- Introduced a new section explaining the optimization process, including layered pipeline steps and measurement tracking.
- Clarified that model usage and companion workflow remain unchanged for users.
v1.0.10
- Removed 3 files: SECURITY.md, persist_spectyra_desktop_config.py, and setup.sh.
- SKILL.md is now more concise, with simplified usage and installation instructions.
- Description and dashboard access details are shorter and more direct.
- Maintains all essential run and model usage information; removed redundant explanations and previews.
- No breaking changes to functionality.
v1.0.9
- Bumped version from 1.0.5 to 1.0.9.
- No other content or functional changes in this release.
v1.0.7
- Removed reference to the dashboard preview screenshot in the SKILL.md
- Slight formatting changes in dashboard preview and section dividers
v9.9.9-testscan
- Updated SKILL.md version to 1.0.6 and clarified that no Spectyra account is required.
- Minor adjustments and clarifications in documentation for setup and usage.
- Maintained instructions and dashboard preview without functional changes.
v1.0.4
Spectyra Skill 1.0.4 Changelog
- Simplified installation and usage instructions in documentation (SKILL.md).
- Updated metadata structure for improved clarity and compatibility.
- No longer requires account or payment to use the companion.
- Instructions highlight single-command setup and easy dashboard access.
- Internal config and metadata JSON files updated for new install method.
v1.0.3
No file changes detected; documentation only update.
- Revised documentation to present setup, usage information more concisely.
- Reorganized sections for clearer install/setup instructions and streamlined dashboard explanation.
- Adjusted technical detail on local companion dashboard URL and reinforced privacy points.
- No changes to code or behavior; this is a documentation/trial/legal copy change only.
v1.0.2
Spectyra 1.0.2
- Added SECURITY.md to clarify security practices, data handling, and address marketplace permission wording.
- Updated trial and payment details in documentation: all upgrades happen on the Spectyra website, not in the skill.
- Expanded Domains & trust section with API domain details and clear statement about no cryptocurrency operations.
- Minor wording updates for accuracy and review clarity.
v1.0.1
- Updated all references from 127.0.0.1 to localhost for clarity and consistency.
- Added a "Domains & trust" section specifying first-party domain ownership and local companion behavior.
- No code or functional changes; documentation improved for transparency and reviewer trust.
- No changes to install, setup, or usage steps.
Metadata
Frequently Asked Questions
What is Spectyra - OpenClaw Workflow & Cost Optimizer?
Optimize OpenClaw workflows and reduce LLM API costs. Runs locally to reduce repeated context, unnecessary steps, and token waste with no workflow changes re... It is an AI Agent Skill for Claude Code / OpenClaw, with 247 downloads so far.
How do I install Spectyra - OpenClaw Workflow & Cost Optimizer?
Run "/install spectyra" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Spectyra - OpenClaw Workflow & Cost Optimizer free?
Yes, Spectyra - OpenClaw Workflow & Cost Optimizer is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Spectyra - OpenClaw Workflow & Cost Optimizer support?
Spectyra - OpenClaw Workflow & Cost Optimizer is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Spectyra - OpenClaw Workflow & Cost Optimizer?
It is built and maintained by ImMrLucky (@immrlucky); the current version is v1.0.25.
More Skills