attio-performance-tuning
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements strong security boundaries by instructing the agent to use existing least-privilege tokens and explicitly forbidding the escalation of scopes for performance measurement tasks.
- [SAFE]: External communication is limited to the official Attio documentation domain (docs.attio.com). The skill uses WebFetch solely for retrieving current API contracts and guidance from this well-known service.
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection as it ingests external data from the provided repository path and telemetry metrics, and possesses the capability to modify files via the Write and Edit tools. However, this functionality is intrinsic to its primary purpose as a performance tuning assistant, and no exploitable patterns or lack of oversight were observed.
- Ingestion points: The [repository-path] argument and telemetry data (latency distributions, query filters) act as inputs to the agent context (defined in SKILL.md).
- Boundary markers: Explicit delimiters for separating untrusted telemetry data from instructions are absent.
- Capability inventory: The skill utilizes Read, Glob, Grep, Write, and Edit tools to analyze and modify the local repository (referenced in SKILL.md).
- Sanitization: No explicit sanitization or validation logic for the ingested telemetry data is defined in the instructions.
Audit Metadata