The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill communicates with Grammarly's official API (api.grammarly.com) via curl and fetch. These network operations are necessary for the skill's primary purpose of diagnosing API connectivity and performance issues.
[DATA_EXPOSURE]: The diagnostic script checks for the presence of environment variables GRAMMARLY_API_KEY and GRAMMARLY_CLIENT_ID. It follows security best practices by masking the actual values in the generated summary logs, only recording whether the variables are set.
[COMMAND_EXECUTION]: The skill provides standard bash scripts for gathering diagnostic information and archiving it into a tarball. The commands used (mkdir, curl, tar, rm) are routine for creating debug bundles and do not involve unauthorized privilege escalation or persistence mechanisms.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes data from Grammarly API responses (text analysis results and account info). While this creates a surface for data processing, the ingestion is limited to well-known service endpoints for specific troubleshooting tasks, representing a low risk profile.

grammarly-debug-bundle