Afterpaths
Give your coding agent searchable memory over past sessions, extracted rules, and what worked so you stop re-discovering the same fixes every sprint.
Overview
Afterpaths is a MCP server for the Operate phase that lets coding agents search past sessions, extract rules, and track what worked across work.
What is this MCP server?
- PyPI afterpaths 0.2.5 via uvx runtime hint; positional entry afterpaths-mcp with --from afterpaths[mcp]
- Search, list, show, and rules tools work without LLM API keys
- Optional ANTHROPIC_API_KEY or OPENAI_API_KEY for session summarization and rule extraction
- Session memory focused on what worked across AI coding agent runs
- Branded Afterpaths / Superjective asset in registry metadata
- Package version 0.2.5
- 2 optional LLM API env vars: ANTHROPIC_API_KEY, OPENAI_API_KEY
- Core tools usable with 0 required secrets for search/list/show/rules per package docs
Community signal: 2 GitHub stars.
What problem does it solve?
Every new agent session starts cold, so you repeat debugging paths and renegotiate project rules you already solved last week.
Who is it for?
Heavy Claude Code or Cursor users who run long iteration loops and want personal session memory without building a custom notes stack.
Skip if: Teams needing centralized wiki governance, HR-safe chat retention, or memory with zero local/Python tooling.
What do I get? / Deliverables
After uvx install and MCP registration, your agent can search prior sessions and reuse extracted rules; optional API keys unlock summarization and rule mining.
- Searchable index of past coding agent sessions
- Persisted rules extracted or curated from prior work
- Trace of approaches that worked for faster iterate loops
Recommended MCP Servers
Journey fit
Operate/iterate is the canonical shelf because session memory pays off when you run production work and repeated agent loops—but the same search applies whenever you code. Iterate subphase matches tracking outcomes, reusing rules, and mining prior sessions rather than one-off codegen.
How it compares
Session-memory MCP for coding agents, not a generic vector database or Git-only audit trail.
Common Questions / FAQ
Who is afterpaths for?
Solo and indie builders using AI coding agents who want to search old sessions, persist rules, and remember effective fixes across projects.
When should I use afterpaths?
Use it whenever you restart agent work—debugging, refactors, launches—and need prior context; summarization features when you set Anthropic or OpenAI keys.
How do I add afterpaths to my agent?
Run via uvx with --from afterpaths[mcp] and positional afterpaths-mcp in MCP config; add ANTHROPIC_API_KEY or OPENAI_API_KEY only if you want summarization and rule extraction.