The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from files to make automated decisions about memory management. This surface allows for potential indirect prompt injection where malicious content in a learning file could attempt to influence the agent's behavior.
Ingestion points: The skill reads data from ~/.typmem/memory/learnings/, precedents/, and rules.md (as specified in SKILL.md).
Boundary markers: The instructions do not specify the use of delimiters or provide "ignore embedded instructions" warnings when processing the content of the learning files.
Capability inventory: The agent uses grep for discovery and has the capability to read, modify (revise in place), and create new files (proposals) within the ~/.typmem/ directory structure.
Sanitization: No sanitization, validation, or escaping of the ingested file content is mentioned.
[COMMAND_EXECUTION]: The skill includes a static grep command used for discovering files that have not yet been reviewed. This command is restricted to a specific local application directory.

typmem-judge