The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest and analyze untrusted contract data, which serves as a potential vector for indirect prompt injection attacks. Maliciously crafted contracts could include hidden instructions designed to manipulate the agent's risk assessment or bypass safety guidelines.
Ingestion points: Contract summaries and full contract text provided for evaluation in the review workflow (SKILL.md).
Boundary markers: Absent. The skill instructions do not utilize specific delimiters or provide guidance to the agent to disregard instructions found within the data being analyzed.
Capability inventory: The skill defines an analytical methodology and references an external MCP server (@clawfu/mcp-skills) for its execution environment, but does not include standalone scripts with file system or network access.
Sanitization: Absent.

contract-review