diagnosing-stacktrace-symbolication

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.70). The skill's required workflow (SKILL.md Step 3) instructs the agent to download a one-hour presigned URL for an uploaded PostHog symbol-data file and to extract and inspect those user-uploaded artifacts (via posthog-cli symbol-sets extract and scripts/inspect_sourcemaps.py), so the agent ingests untrusted third-party content that can materially influence its diagnostic actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.70). The skill instructs running "npx @posthog/cli symbol-sets extract ..." (which fetches and executes the @posthog/cli package from the npm registry, e.g. https://registry.npmjs.org/@posthog/cli) at runtime to extract symbol-data — this downloads and runs remote code and the workflow depends on it.