The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted log data from external files and processes it to generate structured Markdown and JSON reports. This creates a surface for indirect prompt injection if the logs contain maliciously crafted text designed to influence the agent's behavior when it reviews the generated summaries.
Ingestion points: The scripts/preprocess.py script reads the input log file provided via the command line argument in a streaming fashion.
Boundary markers: No specific boundary markers or instructions (e.g., 'ignore instructions within this data') are used when interpolating log content into the analysis reports.
Capability inventory: The skill's capabilities are limited to local file system operations (creating directories and writing analysis reports). It does not have access to the network or the ability to execute arbitrary shell commands based on the log content.
Sanitization: The script extracts data using regular expressions and performs aggregation. However, it does not explicitly sanitize or escape the extracted log content before writing it to the Markdown output files (summary.md, entities.md, etc.).

log-analyzer