The Agent Skills Directory

[SAFE]: The skill provides instructions and examples for the ReasoningBank adaptive learning system, which allows agents to optimize strategies based on historical experience. Analysis of the documentation and code snippets revealed no malicious patterns such as credential theft, remote code execution, or unauthorized network activity.
[PROMPT_INJECTION]: The skill implements a learning loop that ingests external data (task outcomes, patterns, and context), which constitutes an indirect prompt injection surface. 1. Ingestion points: Experience and pattern data enter the system via the recordExperience, learnPattern, and metaLearn methods described in SKILL.md. 2. Boundary markers: The provided implementation examples do not include boundary markers or explicit instructions to ignore potentially malicious embedded content within the ingested data. 3. Capability inventory: The skill's primary logic is focused on strategy recommendations and pattern recognition, and it does not trigger high-risk capabilities like shell execution or file system writes based on the learned data. 4. Sanitization: The snippets do not demonstrate sanitization or validation of the experience data before it is recorded in the persistence layer.

ReasoningBank Intelligence