The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill reads the Sentry authentication token from the user's home directory path ~/.sentryclirc in lib/auth.js. This is the standard location for Sentry CLI credentials.
[DATA_EXFILTRATION]: The authentication token retrieved from the local configuration is transmitted to Sentry's official API endpoint (https://sentry.io/api/0) to authorize data retrieval operations.
[INDIRECT_PROMPT_INJECTION]: The skill fetches and displays untrusted content such as error messages, transaction names, and log entries from the Sentry API, which could potentially contain malicious instructions aimed at the agent.
Ingestion points: Data is ingested through Sentry API responses in all primary scripts (fetch-event.js, fetch-issue.js, list-issues.js, search-events.js, and search-logs.js).
Boundary markers: The skill does not implement specific delimiters or warnings to instruct the agent to ignore potentially malicious content embedded in the fetched Sentry data.
Capability inventory: While the skill scripts are limited to HTTP requests and stdout, the agent using the skill may possess capabilities that could be targeted by instructions found in the ingested logs.
Sanitization: The skill performs no sanitization or filtering of the content returned by the Sentry API before presenting it to the agent.

sentry