fix-posthog-issues
Pass
Audited by Gen Agent Trust Hub on Oct 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATIONPERSISTENCE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from PostHog error logs, including stack traces and event properties. An attacker could trigger an application error with a payload containing instructions to subvert the agent's behavior. Since the agent can modify source code, this presents a risk of unauthorized changes. * Ingestion points: Event samples and stack traces retrieved via PostHog MCP tools in SKILL.md and references/posthog-mcp.md. * Boundary markers: Absent. * Capability inventory: File system write access and task scheduling. * Sanitization: Absent.
- [DATA_EXFILTRATION]: The skill accesses sensitive diagnostic information from PostHog, including session IDs, URLs, and 'person' records containing PII.
- [PERSISTENCE]: The skill uses
CronCreateor ascheduleskill to create future runs for deployment verification, establishing a mechanism for activity across sessions.
Audit Metadata