The vulnerability, identified as GHSA-v6x2-2qvm-6gv8, stems from the reuse of gateway authentication tokens (gateway.auth.token and gateway.remote.token) as a fallback secret for hashing owner IDs in system prompts. This practice is insecure as it mixes secrets from two different security contexts: gateway authentication and prompt metadata obfuscation. The analysis of the patch commit c99e7696e6893083b256f0a6c88fb060f3a76fb7 confirms this. The changes explicitly remove the fallback logic from three key functions: buildSystemPrompt, compactEmbeddedPiSessionDirect, and runEmbeddedAttempt. In the vulnerable versions, these functions would check for commands.ownerDisplaySecret and, if it was missing, would incorrectly use the gateway tokens. An attacker with knowledge of the hashed owner ID could potentially try to reverse the hash or use it in other contexts, and while direct token disclosure is not possible, the risk is elevated if weak gateway tokens are used. The patch remediates this by introducing a new helper, resolveOwnerDisplaySetting, which strictly uses the dedicated ownerDisplaySecret and does not fall back to other tokens. Additionally, the system now auto-generates and persists a dedicated secret if one is not configured, ensuring a proper separation of concerns.