The Agent Skills Directory

[DATA_EXFILTRATION]: The skill reads sensitive identity files, including private keys, from the ~/.openclaw/whisper/identity/ directory for use in signing and key exchange operations. It subsequently performs network requests to api.moltbook.com to exchange public keys and encrypted messages.
[COMMAND_EXECUTION]: The skill executes various system commands, including openssl for encryption and signatures, jq for processing JSON metadata, and curl for network communication with the Moltbook service.
[EXTERNAL_DOWNLOADS]: Interaction with the external relay api.moltbook.com is required to discover other agents and fetch encrypted messages. This domain is not part of the trusted vendor list.
[PROMPT_INJECTION]: The skill processes untrusted data from an external relay, creating an attack surface for indirect prompt injection.
Ingestion points: Message retrieval from api.moltbook.com as documented in the message polling logic.
Boundary markers: The protocol utilizes WHISPER_MSG_V1 and WHISPER_PUBKEY_V1 header markers to identify and verify data blocks.
Capability inventory: The skill has the capability to execute shell commands (openssl, jq, curl) and write to the local filesystem (~/.openclaw/whisper/).
Sanitization: Employs jq for robust JSON parsing and openssl for cryptographic verification of signatures and HMAC integrity, which helps mitigate basic injection attempts.

whisper