feat(core): resilient compaction with reactive retry and middle-out removal

CHANGELOG.md

@@ -7,6 +7,11 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/).
 ## [Unreleased]
 
 ### Added
+- Reactive compaction on `ContextLengthExceeded` — auto-compact and retry LLM calls up to 2 times (#792)
+- `ContextLengthExceeded` error variant in `LlmError` with provider-specific pattern detection (Claude, OpenAI, Ollama)
+- Middle-out progressive tool response removal fallback during summarization (10/20/50/100% tiers)
+- Structured 9-section compaction prompt (User Intent, Technical Concepts, Files & Code, Errors & Fixes, Problem Solving, User Messages, Pending Tasks, Current Work, Next Step)
+- `build_metadata_summary()` — LLM-free final fallback with safe UTF-8 truncation
 - `MessageMetadata` struct in `zeph-llm` with `agent_visible`, `user_visible`, `compacted_at` fields; default is both-visible for backward compat (#M28)
 - `Message.metadata` field with `#[serde(default)]` — existing serialized messages deserialize without change
 - SQLite migration `013_message_metadata.sql` — adds `agent_visible`, `user_visible`, `compacted_at` columns to `messages` table

README.md

@@ -23,7 +23,7 @@ Most AI agent frameworks dump every tool description, skill, and raw output into
 
 - **Semantic skill selection** — embeds skills as vectors, retrieves only top-K relevant per query instead of injecting all
 - **Smart output filtering** — command-aware filters strip 70-99% of noise before context injection; oversized responses offloaded to filesystem
-- **Two-tier context pruning** — selective eviction + adaptive chunked compaction with parallel summarization keeps the window clean
+- **Resilient context compaction** — reactive retry on context overflow, middle-out progressive tool response removal, 9-section structured compaction prompt, LLM-free metadata fallback
 - **Accurate token counting** — tiktoken-based cl100k_base tokenizer with DashMap cache replaces chars/4 heuristic
 - **Proportional budget allocation** — context space distributed by purpose, not arrival order
 
@@ -63,7 +63,7 @@ zeph --tui         # run with TUI dashboard
 |---|---|
 | **Hybrid inference** | Ollama, Claude, OpenAI, Candle (GGUF), any OpenAI-compatible API. Multi-model orchestrator with fallback chains. Response cache with blake3 hashing and TTL |
 | **Skills-first architecture** | YAML+Markdown skill files with semantic matching, self-learning evolution, 4-tier trust model, and compact prompt mode for small-context models |
-| **Semantic memory** | SQLite + Qdrant (or embedded SQLite vector search) with MMR re-ranking, temporal decay scoring, adaptive chunked compaction, durable compaction with message visibility control, credential scrubbing, cross-session recall, vector retrieval, autosave assistant responses, and snapshot export/import |
+| **Semantic memory** | SQLite + Qdrant (or embedded SQLite vector search) with MMR re-ranking, temporal decay scoring, resilient compaction (reactive retry, middle-out tool response removal, 9-section structured prompt, LLM-free fallback), durable compaction with message visibility control, credential scrubbing, cross-session recall, vector retrieval, autosave assistant responses, and snapshot export/import |
 | **Multi-channel I/O** | CLI, Telegram, Discord, Slack, TUI — all with streaming. Vision and speech-to-text input |
 | **Protocols** | MCP client (stdio + HTTP), A2A agent-to-agent communication, ACP server for IDE integration (multi-session, persistence, idle reaper, permission persistence), sub-agent orchestration |
 | **Defense-in-depth** | Shell sandbox, tool permissions, secret redaction, SSRF protection, skill trust quarantine, audit logging |

crates/zeph-core/README.md

@@ -16,15 +16,15 @@ Core orchestration crate for the Zeph agent. Manages the main agent loop, bootst
 | Module | Description |
 |--------|-------------|
 | `agent` | `Agent<C>` — main loop driving inference and tool execution; ToolExecutor erased via `Box<dyn ErasedToolExecutor>`; supports external cancellation via `with_cancel_signal()` |
-| `agent::tool_execution` | Tool call handling, redaction, and result processing |
+| `agent::tool_execution` | Tool call handling, redaction, result processing; `call_llm_with_retry()` / `call_chat_with_tools_retry()` — auto-detect `ContextLengthExceeded`, compact context, and retry (max 2 attempts) |
 | `agent::message_queue` | Message queue management |
 | `agent::builder` | Agent builder API |
 | `agent::commands` | Chat command dispatch (skills, feedback, skill management via `/skill install` and `/skill remove`, sub-agent management via `/agent`, etc.) |
 | `agent::utils` | Shared agent utilities |
 | `bootstrap` | `AppBuilder` — fluent builder for application startup |
 | `channel` | `Channel` trait defining I/O adapters; `LoopbackChannel` / `LoopbackHandle` for headless daemon I/O (`LoopbackHandle` exposes `cancel_signal: Arc<Notify>` for session cancellation); `Attachment` / `AttachmentKind` for multimodal inputs |
 | `config` | TOML config with `ZEPH_*` env overrides; typed `ConfigError` (Io, Parse, Validation, Vault) |
-| `context` | LLM context assembly from history, skills, memory; adaptive chunked compaction with parallel summarization; visibility-aware history loading (agent-only vs user-visible messages); durable compaction via `replace_conversation()`; uses shared `Arc<TokenCounter>` for accurate tiktoken-based budget tracking |
+| `context` | LLM context assembly from history, skills, memory; resilient compaction with reactive context-overflow retry (max 2 attempts), middle-out progressive tool response removal (10/20/50/100% tiers), 9-section structured compaction prompt, LLM-free metadata fallback via `build_metadata_summary()` with safe UTF-8 truncation; parallel chunked summarization; visibility-aware history loading (agent-only vs user-visible messages); durable compaction via `replace_conversation()`; uses shared `Arc<TokenCounter>` for accurate tiktoken-based budget tracking |
 | `cost` | Token cost tracking and budgeting |
 | `daemon` | Background daemon mode with PID file lifecycle (optional feature) |
 | `metrics` | Runtime metrics collection |