The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through its data processing workflows. * Ingestion points: SEC filing section retrieval in workflows/company-profile.md and news/event search in workflows/theme-research.md. * Boundary markers: Absent; the instructions do not provide delimiters or specific commands to ignore instructions found within ingested data. * Capability inventory: All workflows (alert-manager, company-profile, investment-thesis-tracker, theme-research, watchlist-monitor) describe capabilities to create, update, and delete research objects and market alerts within the internal data system. * Sanitization: Absent; no validation or content filtering is specified for the external data before it is presented to the agent.
[NO_CODE]: This skill contains only Markdown documentation and instructions. No executable scripts, binaries, or configuration files are present in the package, which significantly reduces the risk of traditional software vulnerabilities like remote code execution.

llmquant-portfolio