Self-Evolving
v1.0.0Improve reusable agent workflows with reflective experiments, value checks, and local pattern memory.
⭐ 3· 1.3k·9 current·9 all-time
byIván@ivangdavila
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description match the actions described: local experiment logging, promoting proven patterns, and maintaining a small local memory. No unexpected binaries, env vars, or external services are requested.
Instruction Scope
Runtime instructions explicitly confine reads/writes to the active task context and ~/self-evolving/. The skill promises to inform the user before the first persistent write. Minor ambiguity: 'user's main memory' is referenced in setup (what will be saved internally) — this appears to mean the skill's own memory.md but could be interpreted as integrating with a separate memory skill; clarify if the skill will read other skills' storage.
Install Mechanism
No install spec and no code files — instruction-only. This is the lowest-risk install model (nothing is downloaded or executed on install).
Credentials
No required environment variables, credentials, or config paths beyond a benign local directory (~/self-evolving/). The requested access is proportional to the stated feature set.
Persistence & Privilege
always is false and autonomous invocation is allowed (platform default). The only notable persistent privilege is the ability to write local notes in ~/self-evolving/ — this fits the purpose but you should expect the skill to create files unless you opt out.
Assessment
This skill appears internally consistent: it will keep short local notes in ~/self-evolving/ to record experiments and only writes after notifying you. Before installing, consider whether you're comfortable with an agent creating and updating files in your home directory. If you install it: 1) Confirm the first-write notice appears before any files are created. 2) Inspect ~/self-evolving/ (memory.md, experiments.md, archive/) to ensure no sensitive content is stored. 3) Decline or revoke the skill if it later asks for external services, credentials, or permission to read other parts of your system. If you use other 'memory' or persistence skills, clarify how this skill interacts with them (it mentions 'main memory' but doesn't define whether it will read other skills' stores).Like a lobster shell, security has layers — review code before you run it.
latestvk97c61x117z5pymcfn90y0fk9d830p31
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
Runtime requirements
🧬 Clawdis
OSLinux · macOS · Windows