The Agent Skills Directory

Indirect Prompt Injection (LOW): The skill analyzes untrusted documentation files (specs, plans, tasks) which could contain embedded instructions intended to influence the agent's findings or the content of the generated report.\n
Ingestion points: The skill reads spec.md, plan.md, and tasks.md using the Read and Bash (grep) tools.\n
Boundary markers: Missing. There are no explicit delimiters or instructions for the agent to ignore potentially malicious directions embedded within the analyzed documents.\n
Capability inventory: The skill uses the Bash, Read, and Edit tools to process inputs and write an analysis-report.md, creating a path for manipulated data to affect the project state.\n
Sanitization: The skill does not perform any sanitization or validation of the content within the analyzed files before processing it.

analysis-phase