The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests incident timeline data from JSON files which is then processed to generate metrics and formatted reports. This represents a surface for indirect injection if input files contain content designed to manipulate the final report summary.
Ingestion points: The run_analysis function in scripts/agent.py and the main function in scripts/process.py read user-provided incident data files.
Boundary markers: There are no explicit delimiters or instructions defined to separate the untrusted incident data from the report generation logic.
Capability inventory: The skill performs filesystem operations to write JSON and Markdown report files to an output directory.
Sanitization: Input data is processed using standard JSON parsers without secondary filtering or sanitization of text content before inclusion in output reports.
[EXTERNAL_DOWNLOADS]: The skill's documentation and script metadata reference established, well-known Python libraries including requests, jinja2, and matplotlib for installation via standard package managers.

conducting-post-incident-lessons-learned