The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted contract data from external files or user input, which presents a surface for indirect prompt injection attacks.
Ingestion points: The instructions in SKILL.md describe reading documents via the Read tool or capturing pasted text verbatim.
Boundary markers: No specific boundary markers, delimiters, or "ignore embedded instructions" warnings are implemented to separate the contract content from the orchestrator's instructions.
Capability inventory: The skill utilizes Read, Glob, Grep, and the Task tool to perform operations and orchestrate parallel analysis agents.
Sanitization: There is no evidence of sanitization, escaping, or validation of the external contract content before it is interpolated into prompts for the analysis agents.

contract-review