The Agent Skills Directory

[SAFE]: No malicious logic, obfuscation, or unauthorized network operations were detected. The skill's executable scripts are boilerplate scaffolds provided by an enterprise bundler and contain no functional code.\n- [PROMPT_INJECTION]: Indirect prompt injection surface identified. The skill's mandatory Memory Protocol requires the agent to ingest historical context from local files, which may contain untrusted data from previous sessions.\n
Ingestion points: The skill reads from .claude/context/memory/learnings.md at the start of operations.\n
Boundary markers: No explicit delimiters or instructions to ignore embedded commands within the memory files are provided.\n
Capability inventory: The skill uses Read, Write, and Skill tools, enabling data read from memory to influence file system modifications or downstream skill executions.\n
Sanitization: There is no evidence of sanitization or content validation for the data retrieved from the memory directory.

context-degradation