Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Fox Self Evolution
v1.0.0Production-grade autonomous self-improvement system with research-backed meta-learning, safe self-modification, and continuous optimization. Based on AI safe...
⭐ 0· 46·2 current·2 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
The description promises production-grade autonomous self-improvement and explicit ability to modify 'skills and capabilities', 'memory and knowledge', and 'reasoning patterns' without asking. As an instruction-only skill it contains pseudocode describing filesystem-backed modification and backups, yet the registry metadata declares no required binaries, env vars, or config paths. That mismatch (powerful mutating capabilities vs. no declared platform requirements) is incoherent: a skill that truly modifies agent files or other skills would normally need explicit filesystem access instructions or platform-specific privileges.
Instruction Scope
The SKILL.md explicitly authorizes autonomous modification of skills, memory, reasoning, and response formats (it even lists what it CAN modify without asking). While it lists a few things that it MUST ask before doing, the allowed scope is very broad and grants the agent discretion to change other skills and internal state. The instructions include code-like workflows for backing up, applying, testing, and rolling back changes — these imply reading/writing project files, logs, and backups, but there are no concrete constraints or safe-guarding policies tied to the hosting environment.
Install Mechanism
No install spec and no code files (instruction-only), so nothing will be downloaded or written at install time by the registry. This reduces immediate supply-chain risk but does not eliminate runtime risk if the agent is permitted to perform filesystem or network actions later.
Credentials
The skill declares no required environment variables or credentials, yet it instructs the agent to perform sensitive actions that typically require access to filesystem, configuration, or other skills' state. The lack of declared required credentials/config paths is disproportionate to the scope of capabilities the SKILL.md authorizes (self-modification of skills and memory). That gap makes it unclear what privileges will actually be used at runtime and whether sensitive data could be accessed or changed.
Persistence & Privilege
always: false (good) and autonomous invocation is allowed by default. However, because the skill's instructions explicitly permit altering skills, memory, and reasoning without asking, allowing autonomous invocation increases blast radius: an autonomously-invoking agent could implement the documented self-modification flows. The skill does not request persistent installation, but it does request the ability to change persistent state — this combination is risky.
What to consider before installing
This skill lets the agent change its own skills, memory, and reasoning patterns without requiring credentials in the registry metadata — that inconsistency is risky. Before installing, ask: (1) exactly what runtime privileges will the agent have (filesystem write, ability to install packages, edit other skills)? (2) Will human approvals be enforced for any modification to skills, network access, or user data? (3) Can you run it in an isolated/test environment first? If you must try it, restrict the agent's permissions (read-only or sandboxed filesystem), require explicit human confirmation for any modification outside a dedicated test folder, and require code review of any proposed changes. If you cannot confirm these safeguards, do not enable autonomous invocation on production agents.Like a lobster shell, security has layers — review code before you run it.
latestvk9784ta9vz27mjhsjqawmfkevh83t1c3
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
Runtime requirements
🧬 Clawdis
OSmacOS · Linux · Windows