#MODS

Summary: Changes Made to pi-mono

This document outlines features implemented in your pi-mono fork for backport to oh-my-pi:

1. Hashline Editing

• Content-addressable editing using SHA1 hashes (6 hex chars) per line

• Format: LINE#HASH|CONTENT (e.g., 12#a1b2c3|const x = 1;)

• Edit operations: set_line, replace_lines, insert_after

• Hash mismatch recovery: Searches ±8 line window, reports affected ranges for targeted re-read

• Settings toggle: edit.mode = "replace" | "hashline"

• smart partial re-read on mismatch - doesn't flood context with the whole file

2. LSP Integration

• Full LSP client with diagnostics, hover, references, rename, format, symbols
• Modules: client, api, config, installer, settings-state, encounter, planner, edits, render, probe, lspmux
• Per-server settings: Enable/disable, install state, auto-enable/install on encounter
• Agent-guided installation: Detects language, suggests server, provides install command
• TypeScript tuning: Custom rootMarkers, isLinter flag
• Management tools: lspStatus(), lspReload(serverId)

3. ast-grep Support

• AST-based code search (faster than grep for structural queries)
• Tool API: pattern, path, language, include/exclude globs, rule YAML
• First-class built-in tool with auto-install detection

4. Hooks System (Gastown)

• Hook events: SessionStart, PreToolUse, PostToolUse, PostToolUseFailure, PreCompact
• Config sources (precedence): CLI config, project config, Claude settings, Gastown defaults
• Features: Timeout control, output truncation, invocation logging, sensitive value redaction
• Settings toggle: gastownMode (default: true)

5. Capability Policy Externalization

• Skill-based policies moved from hardcoded to external templates
• Two tiers: Core-only (minimal) vs Detailed (full playbook)
• Auto-injection: Detects LSP/ast-grep availability, injects appropriate policy
• Tool selection guidance: LSP vs ast-grep vs grep based on availability
• Multiple refinements: Discovery-first lookup, bounded completeness, lexical backstop requirements

6. Settings & Runtime

• Edit mode toggle: edit.mode = "replace" | "hashline"
• LSP settings: enabled, autoEnable, autoInstall, per-language toggles, per-server controls
• ast-grep settings: enabled toggle
• Live runtime rebuild: Settings changes trigger LSP reinit, tool capability refresh

7. Model Management

• Provider grouping: Models grouped by provider in /model selector
• Free-only filter: defaultModelFreeOnly setting or /freemodel command

8. Additional Features

• Slash command help: /help displays all available commands
• Bundled prompt templates: Load by default with external customization support

9. Current WIP: Smart Partial Re-read

• Commit: 728e30db (in progress)
• Goal: On hash mismatch, only re-read affected line ranges instead of entire file
• Implemented: affectedLineRanges in edit details, ranges param in read tool, mergeRanges() utility
• Missing: Automatic flow to trigger re-read after hash mismatch (3 options: agent-level automation, extension hook, capability policy integration)
• Pending: Handle whitespace lines with repeating hashes

Testing and Issues welcome!

🏖️ OSS Vacation

Issue tracker and PRs reopen February 23, 2026.

All PRs will be auto-closed until then. Approved contributors can submit PRs after vacation without reapproval. For support, join Discord.

---

pi.dev domain graciously donated by


exe.dev

Pi Monorepo

Looking for the pi coding agent? See packages/coding-agent for installation and usage.

Tools for building AI agents and managing LLM deployments.

Packages

Package	Description
@mariozechner/pi-ai	Unified multi-provider LLM API (OpenAI, Anthropic, Google, etc.)
@mariozechner/pi-agent-core	Agent runtime with tool calling and state management
@mariozechner/pi-coding-agent	Interactive coding agent CLI
@mariozechner/pi-mom	Slack bot that delegates messages to the pi coding agent
@mariozechner/pi-tui	Terminal UI library with differential rendering
@mariozechner/pi-web-ui	Web components for AI chat interfaces
@mariozechner/pi-pods	CLI for managing vLLM deployments on GPU pods

Contributing

See CONTRIBUTING.md for contribution guidelines and AGENTS.md for project-specific rules (for both humans and agents).

Development

npm install          # Install all dependencies
npm run build        # Build all packages
npm run check        # Lint, format, and type check
./test.sh            # Run tests (skips LLM-dependent tests without API keys)
./pi-test.sh         # Run pi from sources (must be run from repo root)

Note: npm run check requires npm run build to be run first. The web-ui package uses tsc which needs compiled .d.ts files from dependencies.

License

MIT