Multi Search
v1.1.0Intelligent multi-topic deep research tool supporting arbitrary material input, using independent research agents for parallel deep retrieval and systematic...
⭐ 0· 556·13 current·13 all-time
byNever@neverchenx
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description (multi-topic deep research) match the instructions: parsing input, splitting topics, running per-topic retrieval, and generating files. The lack of required binaries, env vars, or install steps is reasonable for an instruction-only research skill.
Instruction Scope
SKILL.md explicitly directs the agent to read input material (file or pasted text), run multiple rounds of web retrieval (WebSearch/WebFetch), and create output directories/files. This is within the stated purpose, but the skill will write files to the user's filesystem (output/, current dir, or ./research/) and will fetch external web content—users should avoid pasting sensitive material and be aware of where output will be written.
Install Mechanism
No install spec and no code files — instruction-only. This is the lowest install risk and consistent with the described behavior.
Credentials
The skill requests no environment variables, credentials, or config paths. The operations described (reading input, web fetches, writing files) do not require extra secrets, so the absence of requested credentials is proportionate.
Persistence & Privilege
Flags show always:false and user-invocable:true. The skill does not request permanent presence or elevated privileges and does not modify other skills. Autonomous invocation is allowed by platform default but this skill does not escalate that beyond normal.
Assessment
This skill appears to do what it says: accept pasted or file input, run web retrieval rounds, and write structured research files to disk. Before installing/using it: 1) Do not paste or point it at files containing secrets, personal data, or proprietary documents unless you trust the environment—the skill will read those materials and include source content in outputs. 2) Expect it to fetch external web content (WebSearch/WebFetch); if your environment restricts network access, verify how those retrieval methods behave. 3) Check where output will be written (priority: output/[project]/, then ./[project]/, then ./research/) so files are not created in unexpected locations. 4) The skill is instruction-only (no install), so there is no extra software being dropped on disk. Overall there are no unexplained credential requests or hidden endpoints, so the skill is internally coherent and its behavior matches its description.Like a lobster shell, security has layers — review code before you run it.
latestvk97dee4sdsj3sxnjp2nvvh43hs82zcp4
License
MIT-0
Free to use, modify, and redistribute. No attribution required.