Decoupling Prompt Composition from Code
BllaOnline · reddit · 2026-07-13
Tired of repeatedly modifying application code just to tweak prompts, the author developed Peisinoe, a composition library that decouples prompt content and selection logic from business logic.
Core Concepts
- Represents content using reusable prompt components
- Organizes prompt variants via explicit Select, Unit, and Child structures
- Allows prompts to be written in Python or loaded by name from Markdown/YAML .prompt files
Advanced Features
- Added an evaluation layer
- The explicit structure allows the library to enumerate all reachable branches
- Can report which branches lack evaluation coverage (e.g., an untested branch with a free condition)
The author clarifies that this is not a managed prompt management service, nor does it route model calls. It's simply a Python composition library with optional file-based editing and structure-aware evaluation tools.
More from coding & agent
- HeyGen adds a media-sourcing skill for coding agents with 75k images and 10k tracks — HeyGen · 2026-07-22
- Agent search bottlenecks are now about variance, not raw latency — rohanpaul_ai · 2026-07-22
- LangSmith adds tracing for Pipecat, LiveKit, OpenAI Realtime, and Gemini Live — LangChain · 2026-07-22
- An MCP server signs every AI agent tool call into a verifiable Merkle chain — Funky_Chicken_22 · 2026-07-22
- Annotated transcript of a Claude Code team interview is now available — trq212 · 2026-07-22
- Claude Code skill uses 10 Markdown rules to make outputs ADHD-friendly — alex_verem · 2026-07-22