The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill creates a vulnerability surface for indirect prompt injection by directing the agent to ingest and interpret external, untrusted content.
Ingestion points: The instructions in SKILL.md require the agent to read adjacent .c source files and repository metadata files such as AGENTS.md and CLAUDE.md to identify local coding idioms and conventions.
Boundary markers: Absent. The instructions do not provide delimiters or "ignore embedded instructions" warnings for the agent when processing text from external source files or documentation.
Capability inventory: Across the skill and its references (e.g., references/build-and-measurement.md, references/memory-safety.md), the agent is expected to possess and use capabilities for file system access and terminal execution of build and debugging tools like gcc, make, valgrind, and gdb.
Sanitization: Absent. The skill provides no guidance on validating, escaping, or filtering content retrieved from the external environment before it is processed in the agent's logic.

c-systems