The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from the code repository it is indexing, which constitutes an untrusted data source.
Ingestion points: Data enters the agent's context through tools like graft ask (retrieval over the graph), graft grep (regex search over source files), and graft skeleton (parsing file signatures).
Boundary markers: The skill documentation does not specify the use of delimiters or specific instructions to the agent to ignore potentially malicious content embedded within the indexed source files.
Capability inventory: The skill provides the agent with capabilities to read file spans, perform exhaustive searches, and analyze repository-wide call graphs.
Sanitization: There are no mentioned mechanisms for sanitizing or filtering the source code content before it is presented to the agent's context.

graft