The Agent Skills Directory

[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests and processes cluster performance data which could be manipulated to include malicious instructions.
Ingestion points: The scripts scripts/compare_api_stats.py and scripts/compare_op_stats.py read data from user-provided cluster directories via the scripts/rank_data_finder.py utility.
Boundary markers: Absent; there are no delimiters or instructions provided to the agent to disregard instructions that might be embedded within the performance data files.
Capability inventory: The skill is designed to run local Python scripts for data analysis and does not have network access or write permissions to sensitive system files.
Sanitization: No specific sanitization or filtering is performed on the data content before it is processed and presented to the agent context.
[PROMPT_INJECTION]: The skill metadata contains an anomalous synchronization date ('2026-04-18') that is inconsistent with current real-world timelines, representing a form of metadata anomaly.
[COMMAND_EXECUTION]: The skill requires the agent to execute internal Python scripts (compare_op_stats.py, compare_api_stats.py) through the terminal to perform its profiling functions. These scripts are bundled with the skill and do not invoke external network resources.

external-mindstudio-cluster-fast-slow-rank-detector