Skill Manager
v1.0.3Manage installed skills lifecycle: suggest by context, track installations, check updates, and cleanup unused.
⭐ 3· 2.1k·39 current·41 all-time
byIván@ivangdavila
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
medium confidencePurpose & Capability
Name and description match the declared behavior: suggesting skills from context, tracking installs/declines in ~/skill-manager/inventory.md, checking updates, and offering removal. There are no unrelated environment variables, binaries, or config paths requested.
Instruction Scope
SKILL.md clearly limits operations to suggesting based on the current task, asking for explicit consent before install/update/remove, and restricting file activity to ~/skill-manager/. However 'notice the current context' is a behavioral instruction that could be broad in practice — the document claims it will not track usage patterns, but that is a policy-level constraint (not enforced) rather than a technical sandbox.
Install Mechanism
The skill is instruction-only (no install), which is low-risk for this package itself. However its runtime actions use `npx clawhub` to install/update/uninstall other skills — that mechanism downloads and executes code from the ClawHub registry. That behavior is expected for a skill manager but is a meaningful operational risk: installing any third-party skill via npx runs remote code and should be reviewed.
Credentials
No environment variables, credentials, or unusual config paths are requested. The skill stores its inventory in ~/skill-manager/inventory.md and will prompt to store user-declined reasons — this is proportional to its stated purpose, but users should be aware personal notes and skill slugs are written to their home directory.
Persistence & Privilege
The skill is not always-enabled and is user-invocable; model invocation is permitted (platform default). It will write its own inventory under the user's home. Because it can proactively suggest skills (autonomous suggestions), there is some potential for repeated prompts to install remote packages — but the instructions require explicit user consent before executing npx instal/update/uninstall commands.
Assessment
This skill appears to do what it says: suggest skills, keep a small inventory file at ~/skill-manager/inventory.md, and call `npx clawhub` to manage other skills. Before installing or using it, consider: 1) Any installation of other skills via `npx` will download and run code from the ClawHub registry — only install skills you trust and review their source. 2) The skill will store installed and declined skill names (and declined reasons) in your home directory; if that includes sensitive context, avoid recording it. 3) The SKILL.md promises not to read files outside ~/skill-manager/, but that is a behavioral constraint — confirm in practice or limit agent permissions if your environment supports it. If you are uncomfortable with remote code execution via npx or automatic context-based suggestions, do not enable or limit the agent's autonomous invocation.Like a lobster shell, security has layers — review code before you run it.
latestvk97c1ygd9m6489kd9yxpsq643x818y4y
License
MIT-0
Free to use, modify, and redistribute. No attribution required.