claude-capabilities
Pass
Audited by Gen Agent Trust Hub on Feb 18, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE] (SAFE): The skill consists solely of markdown reference files. No scripts (.py, .js, .sh), executables, or automation configurations are present. All logic is descriptive rather than functional.
- [EXTERNAL_DOWNLOADS] (SAFE): The file
references/models.mdincludes examplecurlcommands to query the Anthropic and OpenRouter APIs. These are provided as educational documentation for the user or agent to manually verify model information and are not automatically executed by the skill. - [CREDENTIALS_UNSAFE] (SAFE): Documentation examples in
references/models.mdcorrectly use the placeholder environment variable$ANTHROPIC_API_KEY. No hardcoded secrets or sensitive credentials were found in the provided files.
Audit Metadata