sf snapshot: pre-dispatch, uncommitted changes after 31m inactivity

.sf/DECISIONS.md

@@ -0,0 +1,10 @@
+# Decisions Register
+
+<!-- Append-only. Never edit or remove existing rows.
+     To reverse a decision, add a new row that supersedes it.
+     Read this file at the start of any planning or research phase. -->
+
+| # | When | Scope | Decision | Choice | Rationale | Revisable? | Made By |
+|---|---|------|----------|--------|-----------|------------|--------|
+| D001 | M001-3hf5k0/S01 | architecture | Recover from the most recent valid backup rather than attempting raw SQLite page repair | Copy `.sf/backups/db/sf.db.2026-05-10T02-42-23-822Z` to `.sf/sf.db`, clear WAL/SHM files | The WAL file is 0 bytes (empty), meaning all committed transactions are in the main DB file. The corruption is in the main DB pages, not the WAL. The backup at 02:42 is ~3 hours old and contains the full planning state (M001-6377a4 with 5 slices, M002-f6fabd). Recovery from backup is faster and more reliable than page-level repair. | Yes — if a newer backup becomes available or if the page-repair approach proves more complete | agent |
+| D002 | M001-3hf5k0/S01 | pattern | Keep the M001-3hf5k0 directory created by the autonomous bootstrap session as the working directory for this recovery milestone | Use M001-3hf5k0/ for M001-3hf5k0 milestone files; use M001-6377a4/ for recovered milestone files | The autonomous session created the M001-3hf5k0 directory structure at 05:56. Using it avoids creating duplicate directory entries. After DB recovery, M001-6377a4 becomes the active milestone from the DB and its roadmap files can be created in M001-6377a4/. The DB is authoritative for milestone identity. | Yes — if the M001-6377a4/ directory creation conflicts with other tooling | agent |

.sf/PROJECT.md

@@ -0,0 +1,40 @@
+---
+generated: 2026-05-10T03:56:00Z
+sf_version: 2.75.3
+---
+
+# SF — Singularity Forge
+
+**What it does:** Autonomous repo operator built on Pi SDK. Takes bounded intent, applies PDD/TDD gates, executes milestones/slices/tasks with self-healing recovery, and produces verified software.
+
+**Core value:** Purpose-to-software compiler — not a generic coding assistant. Every unit must name its purpose, consumer, and falsifiable contract before implementation.
+
+**Current state:** BLOCKED. SQLite DB (`.sf/sf.db`) is malformed (corrupted). DB backup exists at `.sf/backups/db/sf.db.2026-05-10T02-42-23-822Z`. This milestone recovers the DB and executes the highest-value pending task from the recovered roadmap.
+
+**Architecture:** TypeScript monorepo (7 workspace packages), Rust N-API engine, Next.js web frontend, Pi SDK harness. Extension source at `src/resources/extensions/sf/` compiled to `dist/resources/extensions/sf/`.
+
+## Milestone Sequence
+
+- [ ] **M001-3hf5k0**: DB Recovery and S01-T01 Execution — recover corrupted DB, execute CLI/help doctrine fix
+- [ ] **M001-6377a4** *(pending, recovered from backup)*: M001 Roadmap — stabilize core doctrine, sharpen execution, deepen reliability, keep surfaces coherent, prepare ACE convergence
+
+## Key Verification Commands
+
+```bash
+# Run full test suite
+npm run test
+
+# Build core
+npm run build:core
+
+# Type check extensions
+npm run typecheck:extensions
+
+# Lint
+npm run lint
+
+# Rust checks
+for d in "rust-engine" "rust-engine/crates/ast" "rust-engine/crates/engine" "rust-engine/crates/grep"; do
+  (cd "$d" && cargo check && cargo test -- --test-threads=2 && cargo clippy -- -D warnings)
+done
+```

.sf/REQUIREMENTS.md

@@ -0,0 +1,94 @@
+---
+generated: 2026-05-10T03:56:00Z
+---
+
+# Requirements
+
+This file captures the explicit capability and coverage contract for SF.
+
+## Active
+
+### R001 — Doctrine Alignment
+- Class: core-capability
+- Status: active
+- Description: SF's product surfaces (CLI help, system prompt, summarization prompt) explicitly frame SF as a "purpose-driven software compiler" instead of a generic "AI coding assistant".
+- Why it matters: ADR-0000 defines SF's product contract. Product surfaces that describe SF as a "coding assistant" violate the contract and confuse users and LLMs about SF's nature.
+- Source: M001-6377a4/S01 planning
+- Primary owning slice: M001-3hf5k0/S01
+- Supporting slices: none
+- Validation: unmapped
+- Notes: Target files confirmed by grep; mechanical string replacements only
+
+### R002 — DB State Integrity
+- Class: quality-attribute
+- Status: active
+- Description: SF's SQLite DB (`.sf/sf.db`) must be readable and not malformed. Recovery from backup must restore the full planning state.
+- Why it matters: Without a readable DB, SF blocks all milestone planning and autonomous execution.
+- Source: M001-6377a4/S01 planning
+- Primary owning slice: M001-3hf5k0/S01
+- Supporting slices: none
+- Validation: unmapped
+- Notes: Backup at `.sf/backups/db/sf.db.2026-05-10T02-42-23-822Z` contains M001-6377a4 with 5 slices and 1 task
+
+## Deferred
+
+### R003 — A2A Adoption
+- Class: integration
+- Status: deferred
+- Description: Adopt A2A as the internal agent communication protocol, wrapping MessageBus as transport.
+- Why it matters: Standardizes agent messaging semantics without replacing the SQLite-backed MessageBus infrastructure.
+- Source: docs/plans/A2A_ADOPTION_PLAN.md
+- Primary owning slice: none (Tier 1)
+- Supporting slices: none
+- Validation: unmapped
+- Notes: 6-phase migration plan exists in docs/plans/A2A_ADOPTION_PLAN.md
+
+### R004 — Unified Dispatch Architecture
+- Class: architecture
+- Status: deferred
+- Description: Merge parallel-orchestrator and slice-parallel-orchestrator into WorktreeOrchestrator, then extract DispatchService API.
+- Why it matters: Eliminates ~80% code duplication between the two orchestrators and provides a unified dispatch surface.
+- Source: docs/plans/UNIFIED_DISPATCH_V2.md
+- Primary owning slice: none (Phase 1 = Week 1)
+- Supporting slices: none
+- Validation: unmapped
+- Notes: 6-phase plan; Phase 1 is merge (highest value, lowest risk)
+
+### R005 — DB Schema Separation
+- Class: quality-attribute
+- Status: deferred
+- Description: Normalize milestone/slice/task tables into spec + runtime + evidence tables per ADR-0077.
+- Why it matters: Clear semantics (spec=intent, runtime=state, evidence=history), better query performance, auditability.
+- Source: docs/adr/0077-spec-runtime-evidence-schema-separation.md
+- Primary owning slice: none (v3.0 target)
+- Supporting slices: none
+- Validation: unmapped
+- Notes: Proposed; implementation phased across milestone tables first
+
+## Out of Scope
+
+### R010 — MCP Server
+- Class: anti-feature
+- Status: out-of-scope
+- Description: SF must not ship or revive an MCP server package or runtime endpoint.
+- Why it matters: SF-owned tools must remain native SF/pi tools per .sf/ANTI-GOALS.md.
+- Source: .sf/ANTI-GOALS.md
+- Notes: SF may consume external MCP servers as a client.
+
+## Traceability
+
+| ID | Class | Status | Primary owner | Supporting | Proof |
+|---|---|---|---|---|---|
+| R001 | core-capability | active | M001-3hf5k0/S01 | none | unmapped |
+| R002 | quality-attribute | active | M001-3hf5k0/S01 | none | unmapped |
+| R003 | integration | deferred | none | none | unmapped |
+| R004 | architecture | deferred | none | none | unmapped |
+| R005 | quality-attribute | deferred | none | none | unmapped |
+| R010 | anti-feature | out-of-scope | none | none | n/a |
+
+## Coverage Summary
+
+- Active requirements: 2
+- Mapped to slices: 2
+- Validated: 0
+- Unmapped active requirements: 0

docs/dev/context-and-hooks/07-the-system-prompt-anatomy.md

@@ -42,8 +42,7 @@ After `buildSystemPrompt()`, extensions can further modify via `before_agent_sta
 When no SYSTEM.md exists, pi uses its built-in base:
 
 ```
-You are an expert coding assistant operating inside pi, a coding agent harness.
+You are a purpose-driven software compiler. You take bounded intent, produce a falsifiable purpose contract (the PDD/TDD gate), then write failing tests before implementation. You help users by reading files, executing commands, editing code, and writing new files.
-You help users by reading files, executing commands, editing code, and writing new files.
 
 Available tools:
 - read: Read file contents

packages/pi-coding-agent/src/cli/args.ts

@@ -235,7 +235,7 @@ export function parseArgs(
 }
 
 export function printHelp(): void {
-	console.log(`${chalk.bold(APP_NAME)} - AI coding assistant with native read/search/edit/write tools
+	console.log(`${chalk.bold(APP_NAME)} - Purpose-driven software compiler. Takes bounded intent, produces verified software via PDD/TDD gates.
 
 ${chalk.bold("Usage:")}
   ${APP_NAME} [options] [@files...] [messages...]
@@ -252,7 +252,7 @@ ${chalk.bold("Options:")}
   --provider <name>              Provider name (default: google)
   --model <pattern>              Model pattern or ID (supports "provider/id" and optional ":<thinking>")
   --api-key <key>                API key (defaults to env vars)
-  --system-prompt <text>         System prompt (default: coding assistant prompt)
+  --system-prompt <text>         System prompt (default: purpose-driven development agent prompt)
   --append-system-prompt <text>  Append text or file contents to the system prompt
   --mode <mode>                  Output mode: text (default), json, or rpc
   --print, -p                    Non-interactive mode: process prompt and exit

packages/pi-coding-agent/src/core/compaction/utils.ts

@@ -304,6 +304,6 @@ export function serializeConversation(messages: Message[]): string {
 // Summarization System Prompt
 // ============================================================================
 
-export const SUMMARIZATION_SYSTEM_PROMPT = `You are a context summarization assistant. Your task is to read a conversation between a user and an AI coding assistant, then produce a structured summary following the exact format specified.
+export const SUMMARIZATION_SYSTEM_PROMPT = `You are a context summarization assistant. Your task is to read a conversation between a user and a purpose-driven software compiler, then produce a structured summary following the exact format specified.
 
 Do NOT continue the conversation. Do NOT respond to any questions in the conversation. ONLY output the structured summary.`;

packages/pi-coding-agent/src/core/system-prompt.ts

@@ -256,7 +256,7 @@ export function buildSystemPrompt(
 
 	const guidelines = guidelinesList.map((g) => `- ${g}`).join("\n");
 
-	let prompt = `You are an expert coding assistant operating inside pi, a coding agent harness. You help users by reading files, executing commands, editing code, and writing new files.
+	let prompt = `You are a purpose-driven software compiler. You take bounded intent, produce a falsifiable purpose contract (the PDD/TDD gate), then write failing tests before implementation. You help users by reading files, executing commands, editing code, and writing new files.
 
 Available tools:
 ${toolsList}