The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill integrates data from the Grok API, which performs live Web and X (formerly Twitter) searches. This data is processed by the agent to generate reports, creating a surface for indirect prompt injection attacks where malicious instructions embedded in web pages or social media posts could influence the agent's behavior.
Ingestion points: External content from web searches and social media feeds is ingested via the research_* functions in scripts/run_research.py and combined with market data.
Boundary markers: No boundary markers, delimiters, or explicit instructions to disregard embedded commands are present in the script or prompt instructions when processing Grok API output.
Capability inventory: The skill utilizes the Bash tool to execute python3 scripts for data processing and analysis.
Sanitization: There is no evidence of sanitization, filtering, or escaping of the external content before it is interpolated into the agent's context for report generation.
[COMMAND_EXECUTION]: The skill executes a local Python script using the Bash tool. The execution instruction in SKILL.md contains a hardcoded absolute path (/Users/kikuchihiroyuki/stock-skills/...) which exposes the local user's account name and filesystem layout.

market-research