The Agent Skills Directory

[SAFE]: The skill implements standard JWT operations using the industry-standard pyjwt library. It includes a robust fallback mechanism for basic decoding using built-in modules when the external library is not present.
[SAFE]: There are no detected network operations, file system modifications, or persistence mechanisms. The tool operates as a stateless command-line utility that outputs results to the console.
[DYNAMIC_EXECUTION]: The script utilizes dynamic imports for the jwt library within functional scopes. This is a legitimate implementation choice to manage optional dependencies and does not pose a security risk in this context.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied JWT tokens. Although this is an ingestion surface for potentially untrusted data, the script is a passive analyzer that does not execute code or perform sensitive operations based on the token's payload content.

jwt-decoder