The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute the shell command date -u +%Y%m%dT%H%M%SZ to generate UTC timestamps for record filenames and log entries.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes user-supplied subjects and facts, which are then used to generate questions and analyze relations.
Ingestion points: User input provided via the <subject> argument and "Given" facts (SKILL.md).
Boundary markers: The skill uses Markdown structural elements (e.g., **Subject:**, 1. [question]) to delimit data, but these do not include explicit instructions for the LLM to ignore potentially malicious embedded commands within that data.
Capability inventory: The skill possesses the capability to read, write, and append to files in the ${CLAUDE_PLUGIN_DATA} and ${user_config.records_mirror} directories, and to execute basic shell commands (SKILL.md).
Sanitization: There is no evidence of sanitization, escaping, or validation of user-provided strings before they are interpolated into the agent's instructions.
[EXTERNAL_DOWNLOADS]: Documentation in the README.md file refers to installation and examples hosted on the author's GitHub repository (github.com/synapseradio/ai-skills). These are recognized as vendor-controlled resources.

connect-ideas