The Agent Skills Directory

[PROMPT_INJECTION]: The skill manages behavioral notes ('instincts') which serve as an indirect prompt injection surface. These markdown files are intended to provide behavioral context but could be used to store malicious instructions.
Ingestion points: Markdown files located in the project-local .claude/instincts/ directory, as specified in SKILL.md.
Boundary markers: Absent; the skill does not define specific delimiters or instructions to prevent the agent from obeying potentially malicious commands embedded within the instinct files.
Capability inventory: The skill has access to Bash, Read, Write, and Glob tools, enabling it to fully manage the file lifecycle on the local system.
Sanitization: Absent; there is no mention of sanitization, filtering, or validation of the content contained within the managed markdown files before they are displayed or acted upon.

instinct-review