podium-contact-dedup

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests contact data from the external Podium API, creating a potential surface where an attacker-controlled record could contain instructions designed to manipulate the agent during the data review process.\n
  • Ingestion points: Contacts are fetched from the Podium API (api.podium.com) via scripts/find_duplicates.py and stored in a local SQLite database.\n
  • Boundary markers: The skill does not provide explicit delimiters or instructions to the agent to ignore potentially malicious text within contact fields when reviewing cluster reports or audit logs.\n
  • Capability inventory: The skill utilizes Write access to the local filesystem, Bash execution for Python and SQLite operations, and network access to interact with the Podium API.\n
  • Sanitization: The implementation uses standard data safety measures such as parameterized SQL queries and JSON serialization, which protect the processing scripts but do not fully isolate the LLM from text-based prompt injection within the data.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 03:48 AM
Security Audit — agent-trust-hub — podium-contact-dedup