The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the shell command mkdir -p memory/ to create a directory structure in the current working directory.
[PROMPT_INJECTION]: The skill generates instructions for the CLAUDE.md file that explicitly instruct the agent to "ignore" the platform's built-in memory system (~/.claude/auto-memory) and prioritize the skill's local files, which is a behavioral override.
[PROMPT_INJECTION]: Indirect Prompt Injection risk factor:
Ingestion points: The skill reads local files such as README.md, package.json, and general code files from the current directory (SKILL.md, Step 1.1) to generate project descriptions.
Boundary markers: No delimiters or specific warnings to ignore instructions within the read files are present when processing this external content.
Capability inventory: The skill has the capability to write files (CLAUDE.md, MEMORY.md, AGENT.md) and create directories (mkdir).
Sanitization: The skill does not perform sanitization or validation of the content read from the local repository before incorporating it into the generated instructions and project summaries.

memory-init