fix(runtime): keep long-running journals and retries reliable

api-surface.json

@@ -940,7 +940,7 @@
       "FileCoordinationLog": "value bd8f02ee8c2c",
       "FileCorpus": "value ac872ff9718b",
       "FileResultBlobStore": "value a464e1a75d77",
-      "FileSpawnJournal": "value aead755d0085",
+      "FileSpawnJournal": "value e9bda7d80f7c",
       "FinalizeContext": "type e45cea6eacef",
       "FinalizerSettled": "type c525a06427fc",
       "FlatWidenGate": "type 825b8b585b72",

docs/agent-managed-compute/reliability.md

@@ -28,6 +28,15 @@ Cancellation and cleanup failures preserve accepted output; budget and execution
 
 The built-in provider recovery path supports one-shot, nonsteering execution.
 Retained execution does not create a steering capability the provider lacks.
+Spawn-journal append validation keeps only rebuildable node, admission, and cursor indexes.
+Healthy appends do not reload prior events; cold reads validate each record once.
+File replacement, truncation, changed metadata, or a failed write invalidates the append index.
+Recovery validates the actual retained bytes before accepting another append.
+The index is not a second durable record or a cross-process ownership fence.
+Observer records detach their inputs before queued I/O, so later hook mutations cannot rewrite the evidence being saved.
+Completed director invocations reset the consecutive transport-failure counter even when the pursuit remains incomplete.
+The explicit total-attempt, deadline, cancellation, and resource bounds still apply.
+
 The file run lock protects one local coordinator.
 It does not fence provider mutations from a partitioned coordinator on another machine.
 No deployed or live multi-provider recovery proof is claimed here.

docs/api/primitive-catalog.md

@@ -7,7 +7,7 @@
 
 # Primitive catalog — the never-stale anti-reinvention inventory
 
-> **GENERATED** from `@tangle-network/agent-runtime@0.231.1` and `@tangle-network/agent-eval@0.182.0` by `scripts/gen-primitive-catalog.mjs`. Do NOT hand-edit — run `pnpm run docs:api`. This is the mechanical companion to the JUDGMENT in `canonical-api.md` (§2 decision table + §1.5 AgentProfile law): that doc says WHICH primitive to reach for and what NOT to build; this catalog proves WHAT exists. Per-symbol signatures + `file:line` live in the per-module pages under `docs/api/`.
+> **GENERATED** from `@tangle-network/agent-runtime@0.232.0` and `@tangle-network/agent-eval@0.182.0` by `scripts/gen-primitive-catalog.mjs`. Do NOT hand-edit — run `pnpm run docs:api`. This is the mechanical companion to the JUDGMENT in `canonical-api.md` (§2 decision table + §1.5 AgentProfile law): that doc says WHICH primitive to reach for and what NOT to build; this catalog proves WHAT exists. Per-symbol signatures + `file:line` live in the per-module pages under `docs/api/`.
 
 ## 1. agent-runtime — own public surface
 

docs/canonical-api.md

@@ -4,7 +4,7 @@
 Generated signatures and the complete export list live in docs/api/.
 Run pnpm docs:freshness after editing this file. -->
 
-> **Version 0.231.1.**
+> **Version 0.232.0.**
 > [`docs/api/primitive-catalog.md`](./api/primitive-catalog.md) lists every export and import path.
 > `agent-eval` must satisfy `>=0.182.0 <0.183.0`.
 > `sandbox` must satisfy `>=0.36.4 <0.41.0`.

package.json

@@ -1,6 +1,6 @@
 {
   "name": "@tangle-network/agent-runtime",
-  "version": "0.231.1",
+  "version": "0.232.0",
   "description": "Shared task-lifecycle skeleton for agents: a recursive loop kernel for chat turns, one-shot tasks, and multi-attempt loops, with trace capture and eval-gated self-improvement. Domain behavior lives in adapters; scoring and ship-gates in @tangle-network/agent-eval.",
   "homepage": "https://github.com/tangle-network/agent-runtime#readme",
   "repository": {

src/durable/observer-journal.ts

@@ -1,6 +1,7 @@
 import { createHash } from 'node:crypto'
 import { readFile } from 'node:fs/promises'
 import { resolve } from 'node:path'
+import { detachedSnapshot } from '../runtime/supervise/snapshot'
 import {
   type RuntimeDecisionPoint,
   type RuntimeHookEvent,
@@ -108,6 +109,15 @@ export class FileObserverJournal implements ObserverJournal {
       )
     }
 
+    // Other hooks receive this same input and may mutate it before queued I/O runs.
+    // Detach now, not after awaiting the previous append or computing the digest.
+    let snapshot: typeof value
+    try {
+      snapshot = detachedSnapshot(value, 'observer record')
+    } catch (error) {
+      return Promise.reject(error)
+    }
+
     let result: ObserverRecord | undefined
     const operation = this.tail.then(async () => {
       this.assertComplete()
@@ -126,8 +136,8 @@ export class FileObserverJournal implements ObserverJournal {
         observedAt: Date.now(),
         ...(this.previousDigest ? { previousDigest: this.previousDigest } : {}),
         ...(kind === 'event'
-          ? { event: value as RuntimeHookEvent }
-          : { decision: value as RuntimeDecisionPoint }),
+          ? { event: snapshot as RuntimeHookEvent }
+          : { decision: snapshot as RuntimeDecisionPoint }),
       }
       const record: ObserverRecord = Object.freeze({
         ...unsigned,