The Agent Skills Directory

[PROMPT_INJECTION]: No malicious injection patterns, safety bypass attempts, or instructions to ignore system guidelines were detected. The instructions follow a professional, technical troubleshooting workflow.
[COMMAND_EXECUTION]: The skill utilizes standard command-line tools such as aio (Adobe I/O CLI), docker, and curl. These are used appropriately for checking service status, retrieving logs, and verifying local endpoint connectivity (http://localhost:4318/v1/traces).
[EXTERNAL_DOWNLOADS]: References to external resources are limited to official Adobe GitHub repositories and OpenTelemetry documentation. These are well-known, trusted sources for technical guidance.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted data, specifically user-provided application code and runtime logs.
Ingestion points: SKILL.md (Phase 2: Get activation logs and Read their code).
Boundary markers: Not explicitly defined in instructions, though typical for interactive troubleshooting sessions.
Capability inventory: The agent has access to shell commands via the aio CLI and diagnostic tools (docker, curl).
Sanitization: The instructions recommend logging sanitized parameters to protect sensitive environment data. The ingestion of logs is necessary for the skill's primary purpose and does not represent a malicious intent.

aio-telemetry-troubleshooting