add-telemetry-instrumentation
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONPRIVILEGE_ESCALATIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests repository code and telemetry patterns using tools like
read_fileandsearch_code(SKILL.md), creating an attack surface where instructions embedded in the analyzed codebase could influence the agent's actions. - Ingestion points: The agent reads existing code to determine logging and tracing patterns.
- Boundary markers: No specific delimiters are mandated for external code interpolation.
- Capability inventory: The skill allows modifying files and executing commands (
write_file,execute_command). - Sanitization: Includes generic instructions to redact credentials from telemetry payloads.
- [PRIVILEGE_ESCALATION]: The skill requires performing kernel-level operations to instrument eBPF socket telemetry (Cilium Hubble, Grafana Beyla), which typically necessitates root or CAP_SYS_ADMIN privileges to load probes into the kernel space.
- [COMMAND_EXECUTION]: Instructions to deploy monitoring agents and configure network-level telemetry involve the execution of system-level commands to modify the runtime environment and install collectors.
- [DYNAMIC_EXECUTION]: The skill incorporates continuous profiling via OTLP and eBPF-based data collection, which involves runtime analysis and dynamic instrumentation of executing application processes.
Audit Metadata