The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted project data to generate summaries, creating a potential surface for indirect prompt injection where malicious content in READMEs or source code could attempt to influence agent behavior during analysis.
Ingestion points: Project files including READMEs, documentation directories (docs/, wiki/), API documentation, configuration files, and source code entry points as defined in SKILL.md.
Boundary markers: Absent; the instructions do not include specific delimiters or instructions to ignore commands embedded within the project files being analyzed.
Capability inventory: The skill uses Glob, Grep, and Read for file analysis and has the capability to write the resulting llms.txt file to the project root.
Sanitization: No explicit sanitization or filtering of ingested file content is specified before the content is processed for summary generation.
[COMMAND_EXECUTION]: The skill utilizes file system traversal and search tools (Glob, Grep, Read) to inspect the project environment. This is the primary and intended functionality for identifying relevant documentation and source files to include in the generated output.

llms-txt