The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from an external source (Yahoo Finance) that could theoretically contain malicious instructions designed to influence the agent's behavior.
Ingestion points: External data is ingested via mcp__chrome-devtools__evaluate_script targeting finance.yahoo.com in SKILL.md.
Boundary markers: None. The skill does not explicitly instruct the agent to ignore instructions embedded in the financial data.
Capability inventory: The skill has access to Bash, Task, Read, and Grep tools, providing a surface for file system interaction and subagent orchestration.
Sanitization: The skill performs sanitization by parsing raw text values into numbers (removing commas, converting to billions/millions), which naturally breaks most text-based injection payloads.

annual-wads