The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill accepts user-supplied text through parameters such as brand, slogan, and hint. This text is rendered into an ASCII art banner. If this generated banner is later read by an LLM in a multi-step process, it could potentially contain malicious instructions. The risk is significantly mitigated by the FIGfont rendering which typically breaks the continuity of characters and words.
Ingestion points: brand, slogan, and hint arguments in scripts/figlet_banner.mjs.
Boundary markers: The script does not wrap the output in specific delimiters or include instructions for downstream agents to ignore the content.
Capability inventory: The script uses figlet for text transformation and process.stdout for output. It does not perform network requests or write to the filesystem.
Sanitization: The script performs basic length limiting and trimming but does not sanitize content for potential prompt injection vectors.
[SAFE]: The skill depends on the figlet npm package, which is a standard library for generating ASCII art. The manual dependency resolution logic in scripts/figlet_banner.mjs is used to facilitate local development and monorepo compatibility, searching within established project paths.

ascii-cli-logo-banner-figletjs