The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill defines a protocol for storing and retrieving arbitrary text notes ('efforts') across sessions. This creates a surface for indirect prompt injection where instructions hidden in stored notes could influence future agent actions.
Ingestion points: The write and update operations in SKILL.md handle external note content for storage.
Boundary markers: The specification does not include requirements for delimiters or instructions to prevent the agent from interpreting stored data as commands.
Capability inventory: The skill is capable of filesystem read/write operations and dynamic loading of other skills.
Sanitization: No sanitization or validation logic is defined for the note content.
[COMMAND_EXECUTION]: The markdown fallback implementation utilizes grep for the find operation, which involves executing shell-level search commands that could be influenced by untrusted 'root' or 'subject' names.
[DYNAMIC_EXECUTION]: The skill supports dynamically loading an 'adapter skill' based on metadata matches in the environment, representing a dynamic loading pattern where third-party code could be invoked at runtime.

effort-store