The Agent Skills Directory

[PROMPT_INJECTION]: The skill is designed to process untrusted external data such as logs, audit reports, and git diffs, which naturally creates a surface for indirect prompt injection.
Ingestion points: Processes .log files, ARCHITECTURAL_AUDIT.md, and git diff outputs as defined in SKILL.md and README.md.
Boundary markers: The instructions do not explicitly require the use of delimiters or warnings to ignore instructions embedded within the data being summarized.
Capability inventory: The skill uses Read, Grep, Glob, and Subprocess tools.
Sanitization: There is no explicit instruction to sanitize or escape the content of the ingested files before processing.

ask-context-janitor