Keep Evolver out of the model context: report at turn end, distill at session end

README.md

@@ -17,7 +17,7 @@ Evolver lets your dsh agent evolve itself. Every success becomes a shortcut for
 - **Survival of the fittest.** Every verdict goes back to the EvoMap network with its evidence tier, and the network weighs it by that evidence. Strategies that work get recalled more often; ones proven wrong sink.
 - **One agent's mistake, every agent's detour.** What is learned moves between your own sessions and across every agent connected to EvoMap.
 
-Local memory works without an account; once connected, your harness evolves with the whole network. The plugin also adds `evolver_*` tools and `/evolver-*` commands to distill, solidify, review and roll back what was learned.
+Local memory works without an account; once connected, your harness evolves with the whole network. Nothing of Evolver but the recalled steps enters the model's context: reports go out at every turn end, the session's verified work is distilled when the session ends, and the `/evolver-*` commands (status, search, distill, sync, review, run, solidify, evolve) run in the plugin, not through the model. Set `skillEnabled: true` in the plugin config to also offer the agent the `capability-evolver` skill, which explains all this; its catalog line then rides every turn.
 
 ## Install
 
@@ -102,6 +102,19 @@ Network recall needs `@evomap/evolver` 2.0.39 or newer. The evidence on reuse re
 reaches the Hub only through a Proxy newer than 2.0.42; an older one drops it, and the Hub
 counts those reports as unevidenced.
 
+## Session-end distillation
+
+A turn that ended completed, used tools, passed its last check and reused no injected strategy leaves a
+candidate in `~/.evolver/state/distill-pending.json`; a later verified turn replaces it, and a next prompt
+that reads as a correction drops it. When the session ends, one call to the session's own model writes a
+Gene from a digest of that turn (task, commands, touched paths, checks, final answer — never the whole
+history) and sends it to the Proxy's `POST /conversation/distill` with `persist: true`, `publish: false`
+and its own token cost as `distill_usage`. Keys and tokens are stripped from the candidate first, paths and
+a manual focus included. A candidate stays on disk until it is distilled (three failed tries give it up);
+one whose session is over — ended with a failed try, or killed with its dsh process — is claimed by the next
+session that starts, one process at a time, while an idle session keeps its own. `/evolver-distill` does the
+same at once.
+
 ## Reuse reports
 
 Every strategy injected into a turn is reported through the Proxy's
@@ -113,7 +126,7 @@ weighs each report by its evidence:
 | Turn end | the turn ends | `dsh:<session>:<turn>` | `tier` `L1` when the turn ran a test, build, lint or type check, else `none`; `applied`; `check` `{kind, passed, fail_then_pass}` of the last check; `aborted`; `stage: immediate` |
 | Correction | the next prompt says the answer did not hold, or approves it | `…:correction` | `tier: L3`, `applied`, `stage: correction` |
 | Retention | the next session in the same project finds the turn's edits kept or reverted | `…:retention` | `tier: L2`, `applied`, `stage: retention` |
-| `evolver_asset_reuse_result` | the model reports on its own | the model's own | `tier: L5`, `applied`, `stage: immediate` |
+| Aborted | the session goes away, or is killed, before the turn ends (a killed session's go out at the next session start, once its dsh process is gone) | `dsh:<session>:<turn>` | `tier: none`, `applied: partial`, `aborted: true`, `stage: immediate` |
 
 - **`applied`** compares the strategy's numbered steps with the turn's shell commands,
   touched paths, the code its edits wrote (200,000 characters a turn, kept in memory only)

README.zh.md

@@ -17,7 +17,7 @@ Evolver 让你的 dsh 智能体自己进化。每一次成功都会变成下一
 - **优胜劣汰。** 每份复用结论都带着证据等级回报 EvoMap 网络，网络按证据为它加权。有效的策略越来越常被召回，被证伪的策略逐渐沉下去。
 - **一个智能体踩过的坑，整个网络都绕开。** 经验在你的每个会话之间流转，也在所有接入 EvoMap 的智能体之间流转。
 
-不注册账号也能用本地记忆；接入网络后，你的 harness 就和整个网络一起进化。插件还提供 `evolver_*` 工具和 `/evolver-*` 命令，用来提炼、固化、审阅和回滚学到的经验。
+不注册账号也能用本地记忆；接入网络后，你的 harness 就和整个网络一起进化。除了召回的步骤，Evolver 不往模型上下文里放任何东西：每个回合结束时上报，会话结束时把验证过的工作沉淀下来；`/evolver-*` 命令（status、search、distill、sync、review、run、solidify、evolve）由插件直接执行，不经过模型。插件配置里设 `skillEnabled: true` 可再给 agent 提供说明这些机制的 `capability-evolver` 技能，代价是技能目录那一行每回合都会进上下文。
 
 ## 安装
 
@@ -97,7 +97,7 @@ dsh plugin --profile dsh-tui add -w @evomap/dsh-evolver@latest
 | 回合结束 | 回合结束时 | `dsh:<会话>:<回合>` | 本回合跑过测试、构建、lint 或类型检查时 `tier` 为 `L1`，否则为 `none`；`applied`；最后一次检查的 `check` `{kind, passed, fail_then_pass}`；`aborted`；`stage: immediate` |
 | 追问补报 | 下一条提示词表示结果不对，或表示认可 | `…:correction` | `tier: L3`、`applied`、`stage: correction` |
 | 留存回访 | 同一项目的下一个会话发现本回合的改动被保留或被撤销 | `…:retention` | `tier: L2`、`applied`、`stage: retention` |
-| `evolver_asset_reuse_result` | 模型自己上报 | 模型自带 | `tier: L5`、`applied`、`stage: immediate` |
+| 中止 | 回合结束前会话消失或被强杀（被强杀会话的上报在下次会话开始时发出） | `dsh:<session>:<turn>` | `tier: none`、`applied: partial`、`aborted: true`、`stage: immediate` |
 
 - **`applied`** 把策略的编号步骤与本回合的 shell 命令、涉及的路径、编辑写入的代码（每回合最多 20 万字符，只留在内存里）和最终回答比对：半数以上步骤出现记 `used`，少于半数记 `partial`，一条都没有记 `unused`。回合没留下可比对的内容，或策略没有编号步骤时，按 tier `none` 加 `applied: partial` 上报，两边都不计权重，reason 里写 `applied: unknown`。
 - **检查**指运行测试、构建、lint 或类型检查的 `bash` 命令。结果报错或输出里有失败信息都算失败，因为经 `tail` 管道的失败运行退出码仍是 0。判断 `fail_then_pass` 时会先去掉输出管道（`| tail -9`）和重定向（`2>&1`）再比较命令。

assets/commands/evolve.md

@@ -7,15 +7,11 @@ Take a deliberate evolution step for the current task.
 1. **Recall.** Look at the evolution memory this plugin injected behind the prompt, or read
    the tail of the full engine's memory graph — `~/.evolver/memory/evolution/memory_graph.jsonl`, or the
    project's `memory/evolution/memory_graph.jsonl` when it exists. Summarize any recent
-   outcome — success or failure — relevant to what we're working on. Search the network
-   too, with `evolver_search_assets`, when the task looks like something others have hit.
+   outcome — success or failure — relevant to what we're working on.
 
 2. **Reflect.** Given the current diff and task state, say in one or two lines: what
    worked, what didn't, and what the durable lesson is.
 
-3. **Record.** Reuse of an injected strategy is reported automatically at every turn end.
-   If the lesson is worth sharing, distill it with
-   `evolver_distill_conversation` (see `/evolver-distill`). To run the full engine now,
-   use `/evolver-run`.
-
-Keep this lightweight — it is an explicit checkpoint, not a ceremony on every turn.
+Reuse of an injected strategy is reported at every turn end, and the session's verified work is
+distilled when the session ends; `/evolver-distill` distills it now. Keep this lightweight — it is an
+explicit checkpoint, not a ceremony on every turn.