The Agent Skills Directory

[PROMPT_INJECTION]: The skill uses a retrieval-augmented approach where stored experience patterns (states, actions, rewards) are used to 'synthesize context' for agent decision-making. This creates a surface for indirect prompt injection if the ingested data contains malicious instructions designed to influence the agent's reasoning.
Ingestion points: The adapter.insertPattern method in SKILL.md stores external data (embeddings and JSON patterns) into a local database.
Boundary markers: Absent; there are no instructions or delimiters provided to ensure the agent distinguishes between stored data and operational instructions during retrieval.
Capability inventory: The skill uses adapter.retrieveWithReasoning and adapter.train to influence agent logic and behavior based on processed data.
Sanitization: Absent; the provided code snippets do not include validation or sanitization of the input data before it is stored or retrieved for synthesis.
[EXTERNAL_DOWNLOADS]: The skill instructs the user to execute npx agentdb@latest, which involves downloading and running executable code from the npm registry at runtime.
[COMMAND_EXECUTION]: The skill utilizes several CLI commands via npx to perform file system operations (creating plugins, managing output directories) and querying project status.

AgentDB Learning Plugins