The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/_cli.py uses subprocess.run to execute the aliyun CLI for performing OpenAPI queries. It also uses subprocess to optionally install the requests package.
[EXTERNAL_DOWNLOADS]: The skill includes functionality to dynamically install the requests Python package from the official registry if the ACTIONTRAIL_AUTO_INSTALL environment variable is enabled.
[DATA_EXFILTRATION]: The skill reads account configuration and credentials from the default vendor path ~/.aliyun/config.json to authenticate its requests. It implements a dedicated masking system to redact AccessKey IDs and account UIDs in all displayed outputs.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes and summarizes external audit log data.
Ingestion points: Audit event data is ingested from the ActionTrail LookupEvents API in scripts/lookup_events.py.
Boundary markers: The output is structured into markdown tables and summaries, but no explicit prompt-level delimiters are used to isolate untrusted content.
Capability inventory: The skill can execute shell commands via subprocess and access local configuration files.
Sanitization: The skill automatically redacts sensitive identifiers (AccessKey IDs, UIDs) using the mask_text and mask_obj functions in scripts/_cli.py before presenting results.

alibabacloud-actiontrail-diagnosis