pi-harness-runtime 0.10.13 → 0.10.15

package/IMPLEMENTATION/RFC-0103/TASKS.md

@@ -1,97 +0,0 @@
-# Tasks — RFC-0103 Session Event Runtime & Agent Communication Layer
-
-## Phase 1 — event-store
-
-- [ ] Create `packages/event-store/src/types.ts` — SessionEvent, SessionEventType, EventMetadata, StoreOptions, StoreStats, WriteResult, ReadOptions, SearchOptions
-- [ ] Create `packages/event-store/src/store.ts` — EventStore class: append(), read(), search(), stats(), list(), getPath()
-- [ ] Create `packages/event-store/src/index.ts` — public barrel
-- [ ] Create `packages/event-store/test/store.test.ts` — bun test with temp dir JSONL
-- [ ] Build and verify: `npx tsc -p packages/event-store/tsconfig.json --skipLibCheck`
-- [ ] Run tests: `bun test packages/event-store/test/store.test.ts`
-
-## Phase 2 — event-bus
-
-- [ ] Create `packages/event-bus/src/types.ts` — Topic union, EventPayload, Subscriber, Subscription, BusOptions, EventBusError
-- [ ] Create `packages/event-bus/src/bus.ts` — EventBus class: publish(), subscribe(), unsubscribe(), unsubscribeAll(), replay(), getSubscribers()
-- [ ] Create `packages/event-bus/src/index.ts` — public barrel
-- [ ] Create `packages/event-bus/test/bus.test.ts` — bun test: pub/sub, priority ordering, filter, unsubscribe, replay
-- [ ] Build and verify
-- [ ] Run tests
-
-## Phase 3 — projection-engine
-
-- [ ] Create `packages/projection-engine/src/types.ts` — Projection<T>, Decision, TaskState, KnowledgeExtraction, AgentTimeline, SessionSummary
-- [ ] Create `packages/projection-engine/src/projections/decision-projector.ts` — detectArchitectureDecision(events): Decision[]
-- [ ] Create `packages/projection-engine/src/projections/task-projector.ts` — detectTaskState(events): TaskState[]
-- [ ] Create `packages/projection-engine/src/projections/summary-projector.ts` — buildSummary(events): SessionSummary
-- [ ] Create `packages/projection-engine/src/projections/timeline-projector.ts` — buildTimeline(events): AgentTimeline[]
-- [ ] Create `packages/projection-engine/src/engine.ts` — ProjectionEngine class: addEvent(), getDecision(), getDecisions(), getTaskState(), getTaskStates(), getSummary(), getTimeline(), reset()
-- [ ] Create `packages/projection-engine/src/index.ts` — public barrel
-- [ ] Create `packages/projection-engine/test/engine.test.ts` — bun test
-- [ ] Build and verify
-- [ ] Run tests
-
-## Phase 4 — workflow-events
-
-- [ ] Create `packages/workflow-events/src/types.ts` — SemanticEvent discriminated union (12 variants), WorkflowState, WorkflowTransition
-- [ ] Create `packages/workflow-events/src/builder.ts` — SemanticEventBuilder: buildFromSessionEvents(), detectTaskAssigned(), detectTaskCompleted(), detectReviewRequested(), detectArchitectureDecision(), detectHumanApproval()
-- [ ] Create `packages/workflow-events/src/workflow.ts` — WorkflowStateMachine: transition(), getState(), getHistory()
-- [ ] Create `packages/workflow-events/src/index.ts` — public barrel
-- [ ] Create `packages/workflow-events/test/workflow.test.ts` — bun test
-- [ ] Build and verify
-- [ ] Run tests
-
-## Phase 5 — subscription-engine
-
-- [ ] Create `packages/subscription-engine/src/types.ts` — Predicate<T>, TopicSubscription<T>, SubscriptionResult, SubscriberInfo
-- [ ] Create `packages/subscription-engine/src/engine.ts` — SubscriptionEngine: subscribe(), unsubscribe(), unsubscribeAgent(), getSubscriptions(), getAgents(), match<T>()
-- [ ] Create `packages/subscription-engine/src/index.ts` — public barrel
-- [ ] Create `packages/subscription-engine/test/engine.test.ts` — bun test: predicate filtering
-- [ ] Build and verify
-- [ ] Run tests
-
-## Phase 6 — session-api
-
-- [ ] Create `packages/session-api/src/types.ts` — SessionService interface, SessionQuery
-- [ ] Create `packages/session-api/src/api.ts` — SessionServiceImpl: getLatestDecision(), getLatestTask(), getRecentFailures(), getArchitectureHistory(), getWorkflowState(), getLatestSummary(), search(), getAgentTimeline(), getSessionTimeline()
-- [ ] Create `packages/session-api/src/index.ts` — public barrel
-- [ ] Create `packages/session-api/test/api.test.ts` — bun test with mock event store + projection engine
-- [ ] Build and verify
-- [ ] Run tests
-
-## Phase 7 — session-export
-
-- [ ] Create `packages/session-export/src/types.ts` — ExportFormat enum, ExportRequest, ExportResult
-- [ ] Create `packages/session-export/src/exporter.ts` — SessionExporter: export(), toMarkdown(), toJson(), toText(), toHtml()
-- [ ] Create `packages/session-export/src/index.ts` — public barrel
-- [ ] Create `packages/session-export/test/export.test.ts` — bun test
-- [ ] Build and verify
-- [ ] Run tests
-
-## Phase 8 — clipboard
-
-- [ ] Create `packages/clipboard/src/types.ts` — ClipboardContent, ClipboardOptions
-- [ ] Create `packages/clipboard/src/clipboard.ts` — copy(), read(), copyLastResponse() with /dev/clipboard fallback
-- [ ] Create `packages/clipboard/src/index.ts` — public barrel
-- [ ] Create `packages/clipboard/test/clipboard.test.ts` — bun test (mock)
-- [ ] Build and verify
-- [ ] Run tests
-
-## Phase 9 — Integration
-
-- [ ] Wire autonomous-runtime (RFC-0101) to publish events via event-bus
-- [ ] Wire context-engineering (RFC-0102) to read session via session-api
-- [ ] Verify all 8 packages build: `npx tsc` across all
-- [ ] Verify all tests pass: `bun test packages/*/test/*.test.ts`
-
-## Phase 10 — Documentation
-
-- [ ] Write `packages/event-store/README.md`
-- [ ] Write `packages/event-bus/README.md`
-- [ ] Write `packages/projection-engine/README.md`
-- [ ] Write `packages/workflow-events/README.md`
-- [ ] Write `packages/subscription-engine/README.md`
-- [ ] Write `packages/session-api/README.md`
-- [ ] Write `packages/session-export/README.md`
-- [ ] Write `packages/clipboard/README.md`
-- [ ] Update root `README.md` with RFC-0103 reference

package/IMPLEMENTATION/RFC-0103/TESTS.md

@@ -1,118 +0,0 @@
-# Tests — RFC-0103 Session Event Runtime
-
-## event-store
-
-### Unit Tests
-
-| Test | Setup | Expected |
-|---|---|---|
-| append generates id | empty temp dir | event.id is valid UUID |
-| append generates timestamp | empty temp dir | event.timestamp is ISO string |
-| append writes JSONL line | empty temp dir | file contains one JSON line |
-| read by sessionId | 2 sessions in dir | returns only matching session |
-| read with since filter | events before/after timestamp | returns only after |
-| read with type filter | message + tool events | returns only matching types |
-| read with limit | 10 events | returns at most 10 |
-| search by content | events with "error" | returns matching events |
-| search query case-insensitive | events with "Error" | returns matching |
-| stats returns count | 5 events | totalEvents = 5 |
-| list sessions | 2 session files | returns 2 IDs |
-| handles corrupt line | JSONL with invalid JSON | skips line, warns |
-
-## event-bus
-
-### Unit Tests
-
-| Test | Setup | Expected |
-|---|---|---|
-| publish returns eventId | subscribe then publish | returns UUID string |
-| subscriber receives event | subscribe + publish | subscriber called with payload |
-| multiple subscribers | 3 subscribers, one topic | all 3 called |
-| unsubscribe removes subscriber | subscribe, unsubscribe, publish | subscriber not called |
-| unsubscribeAll removes all | subscribe to 2 topics, unsubscribeAll | no subscribers |
-| priority ordering | 2 subscribers, priority 1 and 2 | priority 2 called first |
-| filter predicates | subscribe with filter, matching event | subscriber called |
-| filter predicates | subscribe with filter, non-matching event | subscriber NOT called |
-| replay delivers history | subscribe, replay([events]) | subscriber called for each |
-| topic wildcard | subscribe("*") | receives all events |
-| duplicate topic subscribe | subscribe same topic twice | 2 subscriptions, 2 calls |
-| subscriber crash | subscriber throws | other subscribers still called |
-
-## projection-engine
-
-### Unit Tests
-
-| Test | Input | Expected |
-|---|---|---|
-| decision detection | assistant message with "RFC" | Decision.kind = "ArchitectureDecision" |
-| decision detection | assistant message with "decision: use X" | Decision detected |
-| task state detection | tool event with "task-id" | TaskState with correct id |
-| summary counts messages | 5 message events | messageCount = 5 |
-| summary counts tools | 3 tool events | toolCount = 3 |
-| summary infers topics | messages with "deploy" keyword | topics includes "deployment" |
-| timeline groups by agent | events from 2 agents | 2 AgentTimeline entries |
-| reset clears all | add events, reset() | all projections empty |
-| empty events array | [] | graceful, empty results |
-
-## workflow-events
-
-### Unit Tests
-
-| Test | Input | Expected |
-|---|---|---|
-| build TaskAssigned | SessionEvent type="task_assigned" | SemanticEvent.kind = "TaskAssigned" |
-| build TaskCompleted | SessionEvent type="task_completed" | SemanticEvent.kind = "TaskCompleted" |
-| build ReviewRequested | SessionEvent with "review" content | SemanticEvent.kind = "ReviewRequested" |
-| state machine idle->running | TaskStarted event | state.status = "running" |
-| state machine running->finished | WorkflowFinished event | state.status = "finished" |
-| history records transitions | 3 transitions | history.length = 3 |
-
-## subscription-engine
-
-### Unit Tests
-
-| Test | Setup | Expected |
-|---|---|---|
-| subscribe returns id | subscribe agent to topic | returns UUID string |
-| unsubscribe by id | subscribe, unsubscribe | subscription removed |
-| unsubscribeAgent removes all | agent subscribed to 2 topics | all removed |
-| getSubscriptions by agent | 2 agents subscribed | returns only that agent's |
-| match returns predicates | predicate returns true | subscription in match results |
-| match filters | predicate returns false | subscription NOT in match results |
-| priority passed through | subscribe with priority 5 | stored with priority 5 |
-
-## session-api
-
-### Unit Tests
-
-| Test | Setup | Expected |
-|---|---|---|
-| getLatestDecision | 2 decisions in store | returns most recent |
-| getRecentFailures | 3 failure events | returns 3 |
-| search | query "login" | returns matching events |
-| getSummary | mixed events | returns SessionSummary |
-| getAgentTimeline | events from agent-1 | returns AgentTimeline for agent-1 |
-
-## session-export
-
-### Unit Tests
-
-| Test | Input | Expected |
-|---|---|---|
-| toMarkdown has header | events | starts with "# Session" |
-| toMarkdown has timeline | events | contains timestamps |
-| toJson is valid JSON | events | JSON.parse succeeds |
-| toText is plain text | events | no markdown formatting |
-| export with format=markdown | events | content starts with # |
-| export with format=json | events | content is JSON |
-| eventCount in result | 5 events | eventCount = 5 |
-
-## clipboard
-
-### Unit Tests
-
-| Test | Setup | Expected |
-|---|---|---|
-| copy writes text | text "hello" | no error thrown |
-| read returns string | clipboard has text | returns string |
-| copyLastResponse gets latest | sessionApi with messages | latest assistant message copied |

package/IMPLEMENTATION/RFC-0104/ACCEPTANCE_CRITERIA.md

@@ -1,29 +0,0 @@
-# RFC-0104 Acceptance Criteria
-
-## Functional Requirements
-
-- [ ] A2A Adapter sends tasks to remote agents
-- [ ] Agent Card correctly advertises capabilities
-- [ ] Task lifecycle (create, status, cancel) works
-- [ ] Integration with RFC-0103 event bus
-- [ ] Streaming responses via SSE
-- [ ] Push notification support
-
-## Non-Functional Requirements
-
-- [ ] Latency < 500ms for task dispatch
-- [ ] Graceful timeout handling
-- [ ] Retry on transient failures
-- [ ] Authentication on A2A endpoints
-
-## Integration Tests
-
-- [ ] Local task → Remote agent → Response back to local
-- [ ] Task cancellation propagates
-- [ ] Error handling on network failure
-
-## Documentation
-
-- [ ] README with setup instructions
-- [ ] Protocol specification
-- [ ] Example usage

package/IMPLEMENTATION/RFC-0104/FILES.md

@@ -1,33 +0,0 @@
-# RFC-0104 Files
-
-## Source Files
-
-```
-packages/
-  a2a-adapter/
-    src/
-      index.ts           # Main exports
-      agent-card.ts      # AgentCard schema
-      client.ts         # A2A client
-      server.ts         # A2A server
-      types.ts          # Shared types
-```
-
-## Configuration
-
-```
-configs/
-  a2a/
-    agent-card.json      # Local agent manifest
-    server.json         # Server config
-```
-
-## Documentation
-
-```
-docs/
-  a2a/
-    README.md
-    PROTOCOL.md
-    EXAMPLES.md
-```

package/IMPLEMENTATION/RFC-0105/ACCEPTANCE_CRITERIA.md

@@ -1,29 +0,0 @@
-# RFC-0105 Acceptance Criteria
-
-## Functional Requirements
-
-- [ ] TencentDB-Agent-Memory server deployed and accessible
-- [ ] Skills sync from moocoding-skills to server
-- [ ] MCP tools available: tdai_memory_search
-- [ ] Hybrid search (BM25 + vector + RRF) working
-- [ ] OKF format for skill documents
-- [ ] Periodic sync configured
-
-## Non-Functional Requirements
-
-- [ ] Search latency < 200ms
-- [ ] Sync completes in < 5 minutes for full skills folder
-- [ ] Graceful handling of server unavailable
-
-## Integration Tests
-
-- [ ] Agent queries skills via MCP
-- [ ] Skills appear in search results
-- [ ] Sync detects new/updated/deleted skills
-- [ ] Ranked results with scores
-
-## Documentation
-
-- [ ] Server setup guide
-- [ ] MCP tool reference
-- [ ] Sync configuration

package/IMPLEMENTATION/RFC-0105/FILES.md

@@ -1,41 +0,0 @@
-# RFC-0105 Files
-
-## Source Files
-
-```
-packages/
-  knowledge-retrieval/
-    src/
-      index.ts              # Main exports
-      mcp-client.ts        # MCP client for TencentDB
-      sync.ts              # Skills sync pipeline
-      types.ts             # Shared types
-```
-
-## Configuration
-
-```
-configs/
-  knowledge-retrieval/
-    server.json            # TencentDB server URL
-    sync.json             # Sync schedule
-    embeddings.json       # Embedding config
-```
-
-## Scripts
-
-```
-scripts/
-  sync-skills.ts          # Sync skills to server
-  deploy-memory.sh        # Deploy server
-```
-
-## Documentation
-
-```
-docs/
-  knowledge-retrieval/
-    README.md
-    SERVER_SETUP.md
-    MCP_INTEGRATION.md
-```

package/IMPLEMENTATION/RFC-0106/ACCEPTANCE_CRITERIA.md

@@ -1,29 +0,0 @@
-# RFC-0106 Acceptance Criteria
-
-## Functional Requirements
-
-- [ ] SKILL.md files parsed as OKF documents
-- [ ] Knowledge graph with typed nodes (skill, rfc, implementation, concept)
-- [ ] Provenance edges (implements, documents, related_to)
-- [ ] Links to RFC/*and IMPLEMENTATION/* established
-- [ ] MCP tools: kg_query, kg_node, kg_edges, kg_path
-- [ ] Supersession tracking for skill versioning
-
-## Non-Functional Requirements
-
-- [ ] Graph extraction < 10s for full skills folder
-- [ ] Graph query < 100ms
-- [ ] Graph integrity validation
-
-## Integration Tests
-
-- [ ] Agent finds skill by RFC implementation
-- [ ] Graph path from RFC to Skill works
-- [ ] Supersession chain correct
-- [ ] Skills index updated on SKILL.md change
-
-## Documentation
-
-- [ ] Graph schema reference
-- [ ] Provenance mapping guide
-- [ ] Neo4j migration path

package/IMPLEMENTATION/RFC-0106/FILES.md

@@ -1,43 +0,0 @@
-# RFC-0106 Files
-
-## Source Files
-
-```
-packages/
-  knowledge-graph/
-    src/
-      index.ts              # Main exports
-      okf-parser.ts         # Parse SKILL.md to OKF
-      graph-extractor.ts    # Extract nodes and edges
-      provenance-linker.ts  # Link RFC/IMPLEMENTATION
-      mcp-client.ts         # Graph MCP tools
-      types.ts             # Shared types
-```
-
-## Configuration
-
-```
-configs/
-  knowledge-graph/
-    graph.json             # Graph settings
-    extraction.json       # Extraction rules
-    provenance.json       # Provenance mapping
-```
-
-## Scripts
-
-```
-scripts/
-  extract-knowledge-graph.ts  # Build graph from skills
-  migrate-neo4j.ts           # Future: Neo4j migration
-```
-
-## Documentation
-
-```
-docs/
-  knowledge-graph/
-    README.md
-    GRAPH_SCHEMA.md
-    PROVENANCE.md
-```

package/RFC/0001_LOOP_RUNTIME.md

@@ -1,14 +0,0 @@
-# RFC 0001 - Loop Runtime
-
-Purpose: define the autonomous execution loop.
-
-Loop:
-1. Pick task
-2. Assign provider
-3. Code
-4. Test
-5. Review
-6. Repair
-7. Checkpoint
-8. Pause/Resume
-9. Ready for client

package/RFC/0002_PROVIDER_ADAPTER.md

@@ -1,3 +0,0 @@
-# RFC 0002 - Provider Adapter
-
-Defines a provider-neutral interface. Providers implement invoke(), detectQuota(), estimateReset(), normalizeError(), normalizeUsage().

package/RFC/0003_QUOTA_MANAGER.md

@@ -1,3 +0,0 @@
-# RFC 0003 - Quota Manager
-
-Collects quota signals from API responses, provider status, Playwright, and local estimates. Produces provider availability state.

package/RFC/0004_PLAYWRIGHT_ADAPTER.md

@@ -1,3 +0,0 @@
-# RFC 0004 - Playwright Adapter
-
-Uses a persistent browser profile to read provider console information when no official usage API exists. Initial target: MiniMax usage console.

package/RFC/0005_GIT_WORKTREE.md

@@ -1,3 +0,0 @@
-# RFC 0005 - Git Worktree Manager
-
-Allocates one worktree per task, enabling parallel execution, review, resume, and safe merges.

package/RFC/0006_CHECKPOINT_MANAGER.md

@@ -1,56 +0,0 @@
-# RFC 0006 - Checkpoint Manager
-
-## Purpose
-Persist and restore long-running AI coding jobs.
-
-## Motivation
-The runtime must survive provider quota exhaustion, process crashes, network failures, machine restarts, and manual pause without losing progress.
-
-## Goals
-- Save state after every important transition.
-- Restore last known safe state.
-- Support task-level and job-level recovery.
-- Start local-first with JSON files.
-
-## Non-goals
-- Distributed consensus.
-- Multi-user locking.
-- Cloud checkpoint storage.
-
-## Checkpoint Model
-```json
-{
-  "version": 1,
-  "jobId": "job_001",
-  "status": "paused_quota",
-  "requirement": "Build MiniMax quota resume",
-  "currentTaskId": "task_003",
-  "provider": "minimax",
-  "resumeAt": "2026-07-01T20:30:00+07:00",
-  "lastError": "MiniMax 2056 quota exhausted",
-  "createdAt": "2026-07-01T15:20:00+07:00",
-  "updatedAt": "2026-07-01T15:21:00+07:00"
-}
-```
-
-## Storage Layout
-```text
-~/.pi-harness-runtime/jobs/<job_id>/
-  checkpoint.json
-  tasks.json
-  events.jsonl
-  resume_prompt.md
-  last_error.log
-```
-
-## Interface
-```ts
-interface CheckpointManager {
-  save(checkpoint: RuntimeCheckpoint): Promise<void>;
-  load(jobId: string): Promise<RuntimeCheckpoint | null>;
-  appendEvent(jobId: string, event: RuntimeEvent): Promise<void>;
-}
-```
-
-## Recovery Strategy
-Use atomic writes, append-only event logs, schema validation, and human escalation if state cannot be trusted.

package/RFC/0007_SCHEDULER.md

@@ -1,32 +0,0 @@
-# RFC 0007 - Scheduler
-
-## Purpose
-Decide when to run, pause, retry, or resume work.
-
-## Scheduling Reasons
-- quota_reset
-- retry_after_error
-- wait_for_human
-- run_next_task
-- cancelled
-
-## Goals
-- Resume jobs after quota reset.
-- Retry temporary failures with backoff.
-- Persist scheduled resume times.
-- Work locally without cloud services.
-
-## v0.1 Implementation
-Use a local JSON schedule file and polling.
-
-## Interface
-```ts
-interface RuntimeScheduler {
-  scheduleResume(jobId: string, resumeAt: string, reason: string): Promise<void>;
-  dueJobs(now: Date): Promise<string[]>;
-  cancel(jobId: string): Promise<void>;
-}
-```
-
-## Future
-SQLite queue, systemd timer, cron, BullMQ, or external orchestrator.

package/RFC/0008_PROVIDER_ROUTER.md

@@ -1,44 +0,0 @@
-# RFC 0008 - Provider Router
-
-## Purpose
-Select the best provider/model for each task.
-
-## Motivation
-Different providers have different strengths, prices, quotas, and failure modes. The runtime must not hard-code MiniMax, GLM, or Codex into the core loop.
-
-## Provider State
-```text
-available
-limited
-exhausted
-disabled
-unknown
-```
-
-## Routing Inputs
-- task type
-- provider capability
-- quota state
-- previous failures
-- user preference
-- cost policy
-- speed policy
-
-## Example Policy
-```text
-planner -> codex
-code_generation -> minimax
-bug_fix -> glm
-review -> codex
-fallback -> glm
-```
-
-## Interface
-```ts
-interface ProviderRouter {
-  selectProvider(task: RuntimeTask, context: RuntimeContext): Promise<ProviderSelection>;
-}
-```
-
-## Recovery
-If provider fails with quota, mark provider exhausted, ask quota manager for reset time, then schedule resume or select fallback provider.

package/RFC/0009_SHARED_CONTEXT.md

@@ -1,32 +0,0 @@
-# RFC 0009 - Shared Context
-
-## Purpose
-Define how agents share durable context during a long-running coding job.
-
-## Motivation
-A long-running multi-agent coding job needs shared durable state: requirement, plan, task list, decisions, test results, review comments, provider status, and resume prompt.
-
-## Context Layout
-```text
-~/.pi-harness-runtime/jobs/<job_id>/
-  requirement.md
-  plan.md
-  tasks.json
-  decisions.md
-  review.md
-  resume_prompt.md
-  events.jsonl
-```
-
-## Interface
-```ts
-interface SharedContextStore {
-  writeRequirement(jobId: string, text: string): Promise<void>;
-  readRequirement(jobId: string): Promise<string | null>;
-  writeResumePrompt(jobId: string, text: string): Promise<void>;
-  appendDecision(jobId: string, text: string): Promise<void>;
-}
-```
-
-## Design Rule
-Shared context is runtime-owned durable project memory, not one model's chat history.

package/RFC/0010_CONTEXT_WINDOW_MANAGER.md

@@ -1,50 +0,0 @@
-# RFC 0010 - Context Window Manager
-
-## Purpose
-Prevent model context overflow by extracting durable runtime state before the context window becomes full.
-
-## Core Rule
-Chat history is temporary. Runtime context is durable.
-
-## Motivation
-Long-running coding sessions eventually hit context pressure. When context is full, agents forget decisions, repeat work, or require the human to type `resume`.
-
-## Goals
-- Estimate context pressure per provider/model.
-- Trigger compaction before overflow.
-- Extract decisions, task state, open questions, errors, and next action.
-- Write `resume_prompt.md`.
-- Allow a new agent session to continue without human context routing.
-
-## Trigger Policy
-Recommended defaults:
-
-```text
-warning_threshold = 0.60
-compact_threshold = 0.70
-hard_stop_threshold = 0.85
-```
-
-## Artifact Layout
-```text
-harness/context/
-  current_state.md
-  compact_summary.md
-  decisions.md
-  open_questions.md
-  task_progress.json
-  resume_prompt.md
-  last_agent_report.md
-```
-
-## Runtime Events
-- ContextWarning
-- ContextCompactionStarted
-- ContextCompactionCompleted
-- ResumePromptGenerated
-- ContextHardStop
-
-## Acceptance Criteria
-- Runtime can detect context pressure.
-- Runtime can write compact summary and resume prompt.
-- New agent can continue from durable files.