authoring-error-tracking-alerts
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines templates for interpolating untrusted error event data (such as exception names and descriptions) into notifications sent to external platforms.
- Ingestion points: Data enters the system via PostHog error tracking events including
$error_tracking_issue_created,$error_tracking_issue_reopened, and$error_tracking_issue_spiking. - Boundary markers: Templates provided in
references/block-templates.mduse Markdown code blocks and JSON structures to encapsulate untrusted content. - Capability inventory: The skill leverages
posthog:error-tracking-alerts-createto configure automated triggers that send data to Slack, Linear, GitHub, GitLab, and custom Webhooks. - Sanitization: The instructions explicitly mandate wrapping untrusted values in Markdown code blocks to prevent the rendering of malicious links, images, or @mentions. Additionally, it provides guidance on redacting tokens, keys, and identifiers from stack traces and warns against including high-risk fields like request data or local variables.
- [SAFE]: The skill demonstrates strong security awareness by requiring
https://for webhooks and instructing the agent to never echo sensitive information such as Slack workspace IDs or full webhook URLs back to the user.
Audit Metadata