Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Self Evolution
v2.0.0Production-grade autonomous self-improvement system with research-backed meta-learning, safe self-modification, and continuous optimization. Based on AI safe...
⭐ 0· 3k·55 current·60 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 name/description promises an autonomous self‑improvement system and the instructions describe exactly that capability (meta‑learning, safe self‑modification). That alignment is coherent in principle. However, implementing those capabilities legitimately requires elevated access to agent internals, skill files, and persistent storage — access that is not declared or scoped in the package metadata. The absence of declared permissions or config paths is noteworthy.
Instruction Scope
The SKILL.md explicitly authorizes modification of 'skills and capabilities', 'memory and knowledge', and 'system files' (with code snippets showing backup/test/rollback). It also insists the system CAN modify many internal artifacts without asking. These are open‑ended instructions that enable the agent to change other skills and agent state. The document does not enumerate exact files/paths, nor does it provide precise limits, so runtime behavior would depend entirely on platform enforcement — a risky, broad instruction scope.
Install Mechanism
Instruction‑only skill with no install spec and no code files to write or download. From an installation standpoint there is no package download or archive extraction risk.
Credentials
The skill requests no environment variables or credentials, which superficially reduces risk. But the declared abilities (editing skills, system files, memory) implicitly require access to agent configuration, filesystem, or other skills' data — capabilities not declared in requires.env or requires.config. This mismatch between granted authority in text and declared permissions is disproportionate.
Persistence & Privilege
always is false (good), but the skill's runtime instructions authorize modifying other skills and capabilities. That equates to modifying other skills' configuration or agent state, which is specifically flagged as a concern in the evaluation rules. Because the skill tells the agent it 'CAN modify without asking' broad parts of the agent, it effectively requests elevated, persistent privileges even though metadata does not state them.
What to consider before installing
This skill gives the agent broad, high‑privilege authority to self‑modify and to alter other skills and system files, but the package declares no permissions or install steps. Before installing or enabling it: 1) Confirm with the publisher exactly which files, config paths, and internal APIs the skill intends to read or write. 2) Require explicit, auditable confirmation prompts for any change that affects other skills, user data, external communication, or system files. 3) Run the skill only in a strong sandbox or disposable environment until you can verify its behavior and logs. 4) Prefer skills that enumerate precise paths and permission scopes or that avoid self‑modification of other skills. 5) If you cannot obtain clear, narrow permission boundaries, avoid installing this skill in production or on agents with access to sensitive data.Like a lobster shell, security has layers — review code before you run it.
latestvk977t11wxav7ngqdyy6bq9rxtd821cy5
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
Runtime requirements
🧬 Clawdis
OSmacOS · Linux · Windows