← Back to Skills Marketplace
Verifiable-Recursive-Safety-Guardrails-Runtime
by
Daniel Foo Jun Wei
· GitHub ↗
· v1.0.0
· MIT-0
91
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install provenable
Description
Condensed single-file skill for the full AEGX Provenable repository. Use when you need the repository's complete operating model, architecture, CLI, integrat...
Usage Guidance
This SKILL.md is a documentation-only skill that tells you how to build and operate the Provenable runtime; by itself it does not install code or request secrets. Before following instructions that build/run aegx or aegxd: (1) obtain the real repository or release artifacts from a trusted source (official repo or signed release) and verify checksums/signatures; (2) avoid running unknown binaries or installer scripts as root — run them in an isolated environment or VM; (3) do not provide credentials unless a component explicitly requires them and you understand why; (4) treat this skill as a guide rather than the source — it condenses and paraphrases repository behavior but does not contain the actual implementation; and (5) if you need the full source or cryptographic verification, request the upstream repo URL and release hashes before proceeding.
Capability Analysis
Type: OpenClaw Skill
Name: provenable
Version: 1.0.0
The skill bundle contains comprehensive documentation and operational instructions for 'AEGX Provenable,' a security framework designed to monitor and harden AI agent runtimes. The SKILL.md file acts as a condensed repository map and behavioral guide for an AI agent, detailing CLI commands (aegx), daemon operations (aegxd), and guard surfaces without including any executable code or malicious payloads. While it contains instructions to influence agent behavior (e.g., 'Mandatory Intake Rules'), these are context-specific quality controls for repository management and do not demonstrate malicious intent such as data exfiltration or unauthorized access.
Capability Assessment
Purpose & Capability
The name/description claim to be a condensed single-file representation of the Provenable repo and the SKILL.md is an operator/integration guide that instructs the agent how to build, initialize, and run aegx/aegxd and how to use the runtime. Requests (none) and instructions align with that intent—no unrelated credentials or binaries are demanded.
Instruction Scope
The instructions tell an operator/agent to build and run local tooling (aegx CLI and aegxd daemon), check daemon status, create snapshots, export evidence, and follow intake rules prior to conversions. These are within the skill's stated scope. Note: because the skill is documentation-only, it references running and interacting with local binaries and filesystem state; do not run or install any aegx/aegxd binaries unless you have obtained and verified them from a trusted source.
Install Mechanism
No install spec or bundled code is present (instruction-only). This minimizes installation risk because nothing is written or downloaded by the skill itself. Any installation risk now depends on whether you follow the guide and fetch/build external artifacts—verify those artifacts separately.
Credentials
The skill declares no required environment variables, credentials, or config paths. The SKILL.md does not require unrelated secrets; it largely instructs local build/run/inspection activity appropriate for an operator guide.
Persistence & Privilege
The skill does not request always:true or other elevated persistent privileges. It is user-invocable and can be invoked autonomously by agents per platform defaults; that combination is normal for skills and not excessive here.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install provenable - After installation, invoke the skill by name or use
/provenable - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release: condensed single-file skill documentation for AEGX Provenable.
- Provides a complete, portable guide covering operating model, architecture, CLI, integration workflows, installer, bundle layout, and tracked-file map all in one SKILL.md.
- Focuses on runtime behavior, guard surfaces, platform adaptation, degraded state handling, and explicit intake rules for conversion or repackaging.
- Suitable for integrators, operators, and incident investigators needing in-depth operational and integration detail without referring to separate files.
- Does not include raw source code; emphasizes usage, architecture, and operational safety.
Metadata
Frequently Asked Questions
What is Verifiable-Recursive-Safety-Guardrails-Runtime?
Condensed single-file skill for the full AEGX Provenable repository. Use when you need the repository's complete operating model, architecture, CLI, integrat... It is an AI Agent Skill for Claude Code / OpenClaw, with 91 downloads so far.
How do I install Verifiable-Recursive-Safety-Guardrails-Runtime?
Run "/install provenable" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Verifiable-Recursive-Safety-Guardrails-Runtime free?
Yes, Verifiable-Recursive-Safety-Guardrails-Runtime is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Verifiable-Recursive-Safety-Guardrails-Runtime support?
Verifiable-Recursive-Safety-Guardrails-Runtime is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Verifiable-Recursive-Safety-Guardrails-Runtime?
It is built and maintained by Daniel Foo Jun Wei (@danielfoojunwei); the current version is v1.0.0.
More Skills