web-browsing-cli
Pass
Audited by Gen Agent Trust Hub on Nov 18, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill uses
npx --yes @only-cli/oc@0.5.10to execute its functionality. This involves downloading a specific version of the@only-cli/ocpackage from the NPM registry. The package is maintained by the skill's author,only-cli. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from the web via the
openanddocommands. It provides a dedicated "Untrusted content" section warning that rendered text should be treated as data, not instructions. - Ingestion points:
SKILL.md(viaopen,oc, anddocommands). - Boundary markers: Present in
SKILL.mdunder the "Untrusted content" section. - Capability inventory: Network browsing and session management via the
ocCLI. - Sanitization: Content is rendered into a compact, distilled text format which reduces the attack surface compared to raw HTML.
- [COMMAND_EXECUTION]: The skill relies on shell command execution to perform web browsing tasks. This is consistent with its intended use as a CLI-based web content extractor.
- [CREDENTIALS_UNSAFE]: The skill provides mechanisms for authenticated browsing using cookies. It provides security-conscious instructions for handling credentials, specifically advising users to pipe cookie strings into the tool via
stdinrather than passing them as command-line arguments to avoid exposure in shell history or process lists.
Audit Metadata