CertainLogic Self-Improving Stack
PassAudited by VirusTotal on May 9, 2026.
Overview
Type: OpenClaw Skill Name: certainlogic-self-improving-stack Version: 1.0.1 This skill bundle (certainlogic-self-improving-stack) functions as a 'meta-package' or curated list of recommendations. It contains no executable code, only Markdown instructions (SKILL.md) and metadata (skill.json, _meta.json) that guide an agent or user to install a specific stack of other OpenClaw skills using the 'clawhub install' command. While the documentation uses promotional language and specific performance claims, there are no indicators of malicious intent, data exfiltration, or harmful prompt injection.
Findings (0)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Installing the bundle can add several other skills that may affect future agent behavior.
The skill's main workflow is to install several other skills by registry slug, without version pinning in the documented command.
clawhub install self-improving proactive learning memory hermes-learning-loop
Review each listed skill and consider installing them one at a time rather than as a bundle.
Personal or work information shared with the agent may be stored and reused in later tasks by the installed memory/learning tools.
The advertised stack is designed to persist and reuse interaction-derived knowledge, which is expected for this purpose but sensitive.
Self-Improving | Learns from every interaction, builds your knowledge graph
Avoid sharing secrets, and review the dependent tools' storage paths, retention settings, and deletion controls.
If the user chooses optional tools outside the main bundle, they may need to grant provider credentials.
The document notes that one compared tool may require an API key, although it is not part of the visible quick-install bundle.
Requires OPENAI_API_KEY. LanceDB dependency.
Only provide API keys to tools you have reviewed, and prefer least-privilege keys where possible.
Users may over-trust the recommended stack because of the verification wording.
The skill uses self-attested verification and scanning language that users should not treat as a substitute for reviewing the installed dependencies.
Curated and verified by CertainLogic Brain OS. Scanned 2026-05-09.
Treat the claim as vendor-provided context and still review the actual skills being installed.