skills/mukul975/anthropic-cybersecurity-skills/implementing-deception-based-detection-with-canarytoken/Gen Agent Trust Hub
implementing-deception-based-detection-with-canarytoken
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the Thinkst Canary API, which represents a potential surface for indirect prompt injection.
- Ingestion points: Data is ingested via list_tokens and get_alerts in scripts/agent.py from the canary.tools API.
- Boundary markers: The skill does not implement delimiters or ignore instructions to isolate processed API content from the agent's logic.
- Capability inventory: The script can perform network requests to create or delete tokens and acknowledge incidents, and it can read local files for document token creation.
- Sanitization: Results are displayed using JSON serialization but lack specific input sanitization to prevent the interpretation of malicious payloads within the data.
- [DATA_EXFILTRATION]: The skill performs network operations to the Thinkst Canary API at *.canary.tools. This is an established and well-known security service, and the operations are consistent with the skill's stated purpose of managing deception tokens.
Audit Metadata