The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external documentation provided by users for editing and review, creating a potential surface for indirect prompt injection attacks where malicious instructions could be embedded in the processed text.
Ingestion points: Documentation files (READMEs, API docs, guides) and user-provided text snippets are read in Workflow 2 and Workflow 3.
Boundary markers: The instructions do not define specific delimiters or security headers to isolate user-provided content from the agent's instructions.
Capability inventory: The skill uses the agent's standard capabilities for reading, writing, and processing text files. No specialized or high-privilege tools are invoked.
Sanitization: No validation or sanitization is performed on the documentation content before it is processed by the agent.
[SAFE]: The skill consists entirely of Markdown documentation and reference guides. It contains no scripts, binaries, or command execution patterns.
[SAFE]: No network operations, remote downloads, or external URL references (other than placeholders in examples) were detected.
[SAFE]: No obfuscation, hidden characters, or credential exposure were found in any of the analyzed files.

writing-documentation