The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes content from untrusted external sources.
Ingestion points: The skill collects data from social media (Twitter/X), news headlines (Tavily), and financial news via the AIsa API in scripts/trend_forecast.py.
Boundary markers: The synthesis prompt in scripts/trend_forecast.py (SYNTHESIS_PROMPT) lacks explicit boundary markers or instructions for the LLM to ignore potentially malicious commands embedded within the external signals it gathers.
Capability inventory: The skill's configuration in SKILL.md provides the agent with access to the bash, read, and write tools. A successful injection from external data could potentially influence the agent to misuse these capabilities.
Sanitization: There is no evidence of text sanitization or filtering applied to the external data before it is passed to the LLM for synthesis.

trend-forecast