Executable toolbelt for agent self-improvement. Five in-loop primitives — memory, skills, workflow, tasks, and persona — with zero production dependencies, built for Bun.
npm install @tnezdev/spores
# or: bun add @tnezdev/spores# Scaffold .spores/ directory in your project
spores init
# Store a memory
spores memory remember "always emit types from the public API" --tags style,api
# Load a skill
spores skill list
spores skill run release-check | llm
# Track a task
spores task add "update CHANGELOG before tagging"
spores task next
# Activate a persona (pipe into your LLM as system prompt)
spores persona list
spores persona activate spores-maintainer | llm --system -| Primitive | What it does |
|---|---|
| Memory | Tiered store (L1/L2/L3) with recall, reinforce, and dream consolidation |
| Skills | Load .md skill files and pipe their content into an LLM |
| Workflow | Directed-graph runtime — register a graph, create runs, advance node state |
| Tasks | ULID-keyed task queue with status transitions and annotations |
| Persona | Activate a "hat" at the start of a turn: memory tags, skills, task filter, workflow, and a rendered body with live situational facts |
All five primitives read from .spores/ in your project root, with optional global overrides from ~/.spores/.
--json Output as JSON (machine-readable, no truncation)
--wide Disable column truncation in list output
--base-dir Override working directory
The .spores/ directory in this repo is the v0.1 dogfood — a persona, a skill, a workflow, tasks, and memories that are used to build spores itself. Read .spores/README.md for a tour.
See AGENTS.md for the full architecture, on-disk layout, and conventions for agent sessions.