frida
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: No security issues detected. The skill contains high-level instructions for the agent to use the Frida tool for debugging and instrumentation.
- [DYNAMIC_EXECUTION]: The instructions guide the agent to 'Save the narrow hook script' and 'attach to the intended process' to capture runtime events. This behavior is consistent with the primary purpose of the Frida tool and is intended for authorized applications.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow that ingests external data, creating an attack surface for indirect prompt injection.
- Ingestion points: The agent is instructed to 'Read the current scope ledger' in SKILL.md.
- Boundary markers: Absent. No delimiters or warnings to ignore embedded instructions in the ledger are provided.
- Capability inventory: The skill facilitates capabilities including attaching to processes, spawning processes, and executing instrumentation scripts via Frida.
- Sanitization: Absent. There are no instructions for sanitizing or validating the content of the scope ledger before processing.
Audit Metadata