sales-talo
Pass
Audited by Gen Agent Trust Hub on Jul 13, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTION
Full Analysis
- [SAFE]: No malicious patterns or security risks were identified. The skill functions as a legitimate reference for the Talo and Palabra AI platform.
- [EXTERNAL_DOWNLOADS]: The skill provides instructions to install the 'palabra-ai' and 'pip-system-certs' packages from official registries. These are documented as necessary dependencies for the platform's SDKs.
- [REMOTE_CODE_EXECUTION]: Code recipes include curl commands and Python/Node.js SDK usage targeting 'api.palabra.ai'. These are standard API interaction patterns and are not used for executing untrusted remote code.
- [DATA_EXFILTRATION]: No unauthorized data transmission or sensitive file access was found. Authentication examples correctly advise using environment variables for client secrets.
- [INDIRECT_PROMPT_INJECTION]: The skill uses 'references/learnings.md' to store and retrieve accumulated knowledge. 1. Ingestion points: 'references/learnings.md' is read at the beginning of the skill flow. 2. Boundary markers: No delimiters are present to isolate the learnings file content. 3. Capability inventory: The skill can execute shell commands (curl, pip, npm) and write to files. 4. Sanitization: No sanitization of the persisted learnings was detected.
Audit Metadata