LangSkills Search
PassAudited by ClawScan on May 1, 2026.
Overview
This looks like a coherent offline research-search helper, but it asks you to install an external Python package and its retrieved skill text should be verified before use.
Before installing, confirm you trust the langskills-rai package and its source, use an isolated Python environment if possible, and treat retrieved skill bodies as citations or references that need verification rather than instructions to follow automatically.
Findings (2)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Installing it will trust third-party package code and downloaded/local bundles on the user's machine.
The skill is instruction-only but tells the user to install and run an external Python package plus data-bundle installer. This is expected for the search purpose, but the package and bundles are not included in the reviewed artifacts.
pip install langskills-rai # Install matching bundles for the current project or pick a domain: langskills-rai bundle-install --auto
Install only from a trusted source, preferably in a virtual environment, and consider pinning or verifying the package and bundle versions before use.
Search results may be useful evidence, but bad or outdated retrieved text could steer the agent toward unsafe or incorrect actions if copied blindly.
The skill can place full retrieved skill bodies into the agent's working context. That is central to the search function, but retrieved content can influence the agent if treated as authoritative instructions.
`--content` | Include full skill body | off - Use `--content --format markdown` to get copy-paste-ready skill text
Treat retrieved skill text as reference material, verify source URLs and commands before acting on them, and avoid automatically executing copied instructions.