The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill implements a Retrieval-Augmented Generation (RAG) pipeline which creates a potential surface for indirect prompt injection if retrieved documents contain malicious instructions.\n
Ingestion points: Data is ingested through ProcessFileEmbedding and SearchUnified nodes in SKILL.md and retrieved from database collections as shown in references/rag-pipeline.md.\n
Boundary markers: Prompt templates in references/agentic-kit.md use basic text identifiers (Context:, Question:) to segment external data from system instructions.\n
Capability inventory: The skill uses the agentic-kit library to process retrieved context and generate natural language responses.\n
Sanitization: The provided examples do not include explicit sanitization or filtering logic for external content prior to prompt interpolation.\n- [EXTERNAL_DOWNLOADS]: The skill depends on several external Node.js packages for LLM provider abstraction.\n
It utilizes agentic-kit, @agentic-kit/ollama, @agentic-kit/anthropic, and @agentic-kit/openai for model connectivity.\n
It references the official vendor repository at github.com/constructive-io/agentic-kit for library source code.\n- [CREDENTIALS_UNSAFE]: The skill demonstrates secure handling of sensitive authentication data.\n
It instructs developers to use process.env.ANTHROPIC_API_KEY and process.env.OPENAI_API_KEY for API access, which follows secret management best practices by avoiding hardcoded credentials.

constructive-agents