The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides example commands to fetch data from official monitoring platforms including Sentry (sentry.io) and Rollbar (api.rollbar.com). These commands utilize standard API endpoints for error log retrieval.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data in the form of error logs, JSON exports, and CSV files.
Ingestion points: The agent accepts user-provided error logs, JSON exports, and CSV files from Sentry, Datadog, and Rollbar (SKILL.md).
Boundary markers: The instructions do not specify explicit delimiters or "ignore embedded instructions" warnings for the ingested log content.
Capability inventory: The skill uses analysis and reporting capabilities. It also provides examples for external network requests via curl to fetch log data.
Sanitization: The skill includes a normalization step to strip variable parts like IDs and timestamps from error messages, which provides a basic level of data filtering.
[CREDENTIALS_SAFE]: The skill instructions explicitly direct the user to store API tokens in environment variables (e.g., SENTRY_AUTH_TOKEN, ROLLBAR_READ_TOKEN) rather than pasting them into the chat, and mandates that tokens should never be echoed back in reports.

error-monitoring